Arkansas DRG Conversion Plan

Transcription

1 Arkansas DRG Conversion Plan Prepared for: Arkansas Department of Human Services December 29, 2017 navigant.com/healthcare

2 Arkansas DRG Conversion Plan Table of Contents 1 Introduction Evaluating a DRG Method Guiding Principles Basics of a DRG Method DRG Codes and Weights Summary of the DRG Pricing Formulas Basic DRG Pricing Calculation Policy Adjustors Adjustments to DRG Base Transfer Claims Partial Eligibility Provider Loss Outlier s DRG Price versus Final Reimbursement Non-DRG Paid Claims Scope of DRG Method Affected Providers Affected Providers - Discussion Affected Services Affected Services - Discussion Affected Beneficiaries / Medicaid Programs Affected Beneficiaries / Medicaid Programs - Discussion Billing and Review Changes Billing and Review Changes - Discussion DRG Grouping DRG Grouper DRG Grouper - Discussion DRG Grouper - Recommendation DRG Relative Weights DRG Relative Weights - Discussion DRG Relative Weights - Recommendation Provider Base Rates Provider Base Rate Categories Provider Base Rate Categories - Discussion Standardized Base Rates with Wage Area Adjustments Standardized Base Rates with Wage Area Adjustments - Discussion...26 Page i December 29, 2017

3 Arkansas DRG Conversion Plan 6.3 Funding for Provider Base Rates Funding for Provider Base Rates - Discussion Per Diem Base Rates Per Diem Base Rate - Discussion Pricing Logic Pricing Flow Policy Adjustors Policy Adjustors - Discussion Transfer Adjustments Transfer Adjustments - Discussion of Non-General Revenue Funds Distributed on a Claim-by-Claim Basis of Non-General Revenue Funds Distributed on a Claim-by-Claim Basis - Discussion Other Per Claim Add-On s Other Per Claim Add-On s - Discussion Outlier s Outlier s - Discussion Provider Gain Adjustments Provider Gain Adjustments - Discussion Short Stay Adjustments Short Stay Adjustments - Discussion Non-Covered Days Adjustments Non-Covered Days Adjustments - Discussion Other Considerations Admit versus Discharge Date Admit versus Discharge Date - Discussion Transitional Period Transitional Period - Discussion Documentation and Coding Adjustment Documentation and Coding Adjustment - Discussion Interim Claims and Late Charges Interim Claims and Late Charges - Discussion Medicare Crossover Comparison Pricing Medicare Crossover Comparison Pricing - Discussion DRG Simulation Model DRG Model Analytical Dataset DRG Model Analytical Dataset - Discussion...45 Page ii December 29, 2017

4 Arkansas DRG Conversion Plan 9.2 Modeling Scenarios Modeling Scenarios - Discussion Modeling Findings Modeling Findings - Discussion DRG Implementation Task 1 Finalize APR DRG Method Design Task 2 Evaluate Assessment and IGT Arrangements Task 3 Stakeholder Engagement and Training Task 4 Develop DRG Calculator Task 5 - Administrative Code/State Plan Changes and UPL Task 6 Develop Access Monitoring Review Plan Task 7 Determine MMIS Business Requirements and Make Required System Changes...54 Appendix A: Summary Results of DRG Modelling for Arkansas Appendix B: Arkansas Hospital Association Response Appendix C: Supporting Data CMS Wage Index by CBSA (FY18) - Arkansas County to Region Assignment Page iii December 29, 2017

5 Arkansas DRG Conversion Plan 1 Introduction This report provides a Conversion Plan to the Arkansas Department of Human Services (DHS) for converting Arkansas Medicaid s inpatient hospital reimbursement methodology from the current per diem-based approach to a Diagnosis Related Group (DRG) methodology. This report in response to the Arkansas legislature, which in House Bill 1016 required that DHS, in coordination with the Arkansas Hospital Association, develop a plan for an inpatient DRG conversion by December 31, The mandated Conversion Plan included the following requirements: How supplemental payments to hospitals will be considered Whether funding for the transition to APR DRGs will be provided to hospitals Whether certain hospitals will be exempt from APR DRGs Estimated impacts of conversion for both general acute hospitals and CAHs DHS engaged Navigant Consulting, Inc. (Navigant) to develop this Conversion Plan, conduct APR DRG payment simulation modeling and describe model findings, methodology options and implementation steps for consideration. It is important to note that payment simulation modeling, rates and other analysis conducted for this report are for the purposes of evaluating the feasibility of converting to a DRG system, and do not represent final rates, recommendations or decisions made by DHS. This report describes which types of providers and services might be included in a change from the current per diem payment method to DRG payment. Also, included is a description of how DRG pricing differs from per diem pricing along with discussion of the various payment method design options applicable when using DRGs as a basis for reimbursement. Each design option has an impact on reimbursement to individual hospitals, and decisions on a full set of design options is required in order to estimate change in reimbursement for individual hospitals resulting from an implementation of DRG pricing. DRG-based inpatient payment systems are the predominate methodology used by State Medicaid agencies nationally. Pricing for hospital inpatient services via DRGs provides a better correlation than per diem payments between the level of resources expended by the hospital when treating a patient and the reimbursement provided by Medicaid. DRG classifications also allow for easier tracking of the types of services being reimbursed by Medicaid and provide a foundation that can be used for measuring hospital quality of care and case mix adjusting those measures to allow for comparison across hospitals. DRG pricing may also provide an avenue through which some or all hospital supplemental payment funding can be moved into claimbased payment rates for inpatient admissions in preparation for the transition of traditional feefor-service (FFS) populations into managed care, such as the implementation of the Provider- Led Arkansas Shared Savings Entity (PASSE). Specifically, the first two chapters of the document provide background on DRG pricing that is helpful in evaluating the various pricing design considerations. Chapter 1 lists a series of criteria helpful in evaluating any Medicaid payment method and describes some of the areas in which options in a DRG pricing method affect the criteria. Chapter 2 describes the components of a standard DRG pricing calculation, including a few optional components, such as policy adjustors. Chapters 3 through 8 provide a comprehensive list of options available to customize a DRG pricing method considering the experience of other state Medicaid agencies and Page 1 December 29, 2017

6 Arkansas DRG Conversion Plan Medicare. Chapter 9 describes the payment simulation model approach and model findings and Chapter 10 provides an overview of DRG implementation tasks and estimated time requirements. Appendix A includes summaries for three different payment simulation model versions. Appendix B includes a response from the Arkansas Hospital Association, based on their review of a draft of this report and a hospital stakeholder meeting conducted on December 12, Appendix C includes reference data used in report. For each DRG payment design option described in this document, a discussion is provided that explains how the option will affect pricing and trade-offs of various choices. If the Arkansas Legislature and DHS decide to move forward with DRG pricing, then additional consideration and stakeholder discussions will be required to finalize the set of options that will comprise the full DRG pricing method for Arkansas Medicaid. In this report we have included recommendations for two major options for which we believe are the most appropriate for Arkansas Medicaid. Specifically we recommend use of All-Patient Refined Diagnosis Related Groups, or APR DRGs (as specified in the Legislation mandating this study) and the use of 3M APR DRG standard national weights. All other system parameters will require additional modeling, stakeholder input and DHS decision making before finalizing. In Chapter 9 and in Appendix A, we provide results of three models in which we have selected a set of payment parameters and simulated DRG payments using these parameters. These models are intended as examples of how various design options and associated payment parameters can affect reimbursement for individual hospitals. We expect this information will be provided by DHS to the Arkansas Legislature, the Arkansas Hospital Association, and other interested stakeholders in early January Further stakeholder meetings and pricing modeling beyond that presented in this document will likely be needed to ensure that the final DRG payment method best meets the needs of the Arkansas Legislature, DHS, and the hospitals providing care for Arkansas Medicaid recipients. In addition, we aim to create a design that is sufficiently flexible to react to any changes required in future years through changes in configuration data, such as rates and policy adjustors, without requiring any additional changes to the software used to adjudicate claims beyond what we have already defined. Page 2 December 29, 2017

7 Arkansas DRG Conversion Plan 2 Evaluating a DRG Method Guiding Principles Developing a Medicaid payment method requires balancing a variety of trade-offs and competing priorities, given limited available system funding. methods have an impact on beneficiaries, medical providers, taxpayers, and program administrators, each with their own point of view on what makes a payment method successful. To balance the priorities of these different stakeholders, it is helpful to establish a set of guiding principles that describe the goals of the payment method and offer a structure against which various system design options can be evaluated. The list below offers a series of guiding principles and discusses how these principles can affect a DRG payment method. Efficiency. A payment method should be consistent with incentivizing hospital efficiency, rewarding hospitals that increase efficiency while continuing to provide quality care. To enable this, the payment method should minimize reliance on individual hospital charges or costs, and create opportunities for providers to increase margins by more effectively managing resources. For example, in the design of a DRG payment system, selecting a single standardized base rate can create incentives for hospitals to better manage their resources to achieve improved margins. Conversely, establishing facility-specific base rates that fluctuate annually with increases or decreases in facilityspecific costs would provide little incentive for cost effectiveness. Access. A payment method should promote beneficiary access to care. This guiding principle is consistent with the requirements specified in federal regulation. In the State Plan for Medical Assistance (State Plan), DHS must make certain assurances to the Federal Centers for Medicare and Medicaid Services (CMS) with respect to its level of payments to Medicaid providers. In particular, the State Plan must: provide such methods and procedures relating to the utilization of, and the payment for, care and services available under the plan as may be necessary to safeguard against unnecessary utilization of such care and services and to assure that payments are consistent with efficiency, economy, and quality of care and are sufficient to enlist enough providers so that care and services are available under the plan at least to the extent that such care and services are available to the general population in the geographic area[.] 42 U.S.C. 1396a(a)(30)(A) ( Section 30(A) ) (emphasis added). Within a DRG payment method, policy adjustors, provider peer groups (used for setting base rates), and outlier payment parameters are items that can be adjusted to affect access to care. Equity. A payment method should generate fair payments across both hospitals and types of care. Generally, hospitals should be paid similar amounts for the same services, with the potential exception being when there are necessary and measurable differences in the costs associated with those similar services. Within a DRG payment method, the bulk of the payment amount for an individual hospital stay is calculated by multiplying a hospital base price times a DRG relative weight. The DRG relative weights are determined using average costs from many hospitals, so the relative weights help ensure similar payment for similar services, independent of where those services are provided. If adjustments do need to be made for reasonable, measurable differences in hospital cost structures, those can be made through modifications to the hospital base Page 3 December 29, 2017

8 Arkansas DRG Conversion Plan price via rate adjustments (for example, wage area adjustments) and/or provider peer groupings (for example, giving all children s hospitals or all rural hospitals their own provider base rate). Predictability. A payment method should generate stable, predictable payments. Both the state Medicaid agency and the hospitals have to manage their budgets, and that can best be facilitated through a payment method which generates consistent, predictable reimbursements. DRG payment methods are predictable if patient acuity and volume are understood. Transparency. A payment method that is transparent promotes trust from hospital administrators, hospital clinicians, legislators, and Medicaid program administrators. A DRG payment method can be made transparent by selecting a DRG algorithm that is openly documented, and by making DRG relative weights, provider base rates, and pricing logic publicly available. Simplicity. A payment method that is relatively simple will be easier to implement, easier for hospitals to understand, and easier to administer and maintain. For a Medicaid program, implementing a new DRG payment method will require significant Medicaid Management Information System (MMIS) changes, regulation changes, and program monitoring changes. For hospitals, a new DRG payment method may impact medical coding practices, billing procedures, and internal information systems. The complexity of these changes is limited if the payment method is kept relatively simple. At the same time, over-simplifying the payment method may negatively impact payment equity and, in turn, negatively impact access to care. Quality. It is generally known that it is a mission of all hospitals to provide high quality care. methods should be consistent with promoting quality care where possible. In truth, very few payment methods specifically reward quality. Most payment methods, including DRG payment methods, pay the same without regard to the provision of high quality care being provided. However, DRG payment methods categorize the complexity of the patient s illness as standard part of processing. This allows for case mix-based risk adjustment of quality measures which enables more fair comparison between hospitals that treat the most complex cases, versus those that treat primarily routine illnesses. From a logistical point of view, a payment method is a framework or structure created to determine reimbursement for medical services and supplies. The structure includes organization of data, numerical formulas, and specific parameters or values used in the formulas. This structure should be carefully developed as it controls the distribution of large amounts of state and federal funding, and is intended to meet the needs of people and organizations with competing priorities. The guiding principles presented above can be helpful in evaluating various options for the payment structure so that the final design best meets the needs of beneficiaries, providers, taxpayers and program administrators. Page 4 December 29, 2017

9 Arkansas DRG Conversion Plan 3 Basics of a DRG Method This section describes the calculations performed in a typical DRG payment method. Ultimately, a payment method can be described as a series of calculations. As such, this section offers a context for how decisions on the various pricing options are applied to actually price claims. Discussions of each component within these calculations are provided in Chapter DRG Codes and Weights DRG payment methods involve classifying inpatient stays and then determining a price based on a combination of the classification and the hospital where the services were performed. Classification of the hospital stay is based on the diagnoses describing the patient s condition, the surgical procedures performed (if any), patient age, and discharge status. The classifications are labeled using codes referred to as DRG codes and the number of codes varies depending on the selected patient classification model. For example, the MS-DRG grouping method has 752 total valid codes, including base codes separated by severity into no CC, with CC or with major CC (where CC stands for complications and comorbidities). Similarly, the APR DRG grouping method has 1,304 codes including 326 base codes each separated into four levels of Severity of Illness (SOI) levels: 1 - minor, 2- moderate, 3-major and 4-extreme. Each DRG code is assigned a relative weight which is intended to indicate the average relative amount of hospital resources required to treat patients within that DRG category. These weights are relative to the overall average amount of hospital resources needed to treat a patient when looking across the full range of patients treated within an acute care inpatient setting. For example, a DRG weight of 2.0 would indicate an admission that requires twice the level of resources as an average admission, while a DRG weight of 0.5 would indicate an admission that requires half the level of resources as an average admission. 3.2 Summary of the DRG Pricing Formulas A summary of a typical DRG pricing calculation is shown in Figure 3.2 and the formulas are described in more detail in the following sections. Page 5 December 29, 2017

10 Arkansas DRG Conversion Plan Figure 3.2 Typical DRG Formulas 1) [Full DRG base pymt] = [Hospital base rate] * [DRG rel wt] * [Policy adjustor(s)] 5) [Estimated cost] = [Covered charge] * [Hospital cost-to-charge ratio] 2) If transfer, [per diem amt] = {[DRG base pymt] / [DRG avg LOS]} * (LOS + 1) 6) [Estimated gain/loss] = AbsVal{[Estimated cost] - [DRG base pymt]} 3) If partial elig, [per diem amt] = {[DRG base pymt] / [DRG avg LOS]} * (LOS + 1) 7) If [Estimated gain/loss] > outlier threshold then outlier payment applies 4) If transfer or partial elig, 8) If hospital loss, [DRG base pymt] = lessor of [Full DRG base pymt] and [per diem amt] [Outlier pymt] = [Estimated gain/loss] * [Marginal cost percentage] Else Else [DRG base pymt] = [Full DRG base pymt] [Outlier pymt] = [Estimated gain/loss] * [Marginal cost percentage] * -1 9) [DRG allowed amount] = [DRG base pymt] + [Outlier pymt] 10) [Reimbursement amount] = [DRG allowed amount] - [Other ins pymt] - [Spend down] - [Cost sharing] Notes: Formulas are typical and can be modified to meet a state's specific needs. "pymt" is an abbreviation for "payment". "LOS" is an acronym for "length of stay". 3.3 Basic DRG Pricing Calculation In a DRG pricing method, the vast majority of hospital stays are priced using a very simple formula. The formula is: [DRG Base ] = [Hospital Base Rate] * [DRG Relative Weight] * [Policy Adjustor(s)] Policy adjustors, which are discussed in the next section, are optional and in many cases and are set to 1.0, indicating no adjustment. If a policy adjustor of 1.0 is assumed, an example claim from a provider with a DRG base rate of $7,000 and a DRG with relative weight of 2.0 would yield a payment of $14,000. Similarly, an admission to the same provider that gets assigned a DRG with relative weight of 0.5 would yield a payment of $3,500. Although this calculation is quite simple, a great deal of effort goes into development of the DRG grouping algorithm (which determines the DRG code), assignment of relative weights to DRG codes, and assignment of base prices to hospitals. 3.4 Policy Adjustors Medicaid agencies can make a policy decision to increase (or decrease) payments for particular types of hospital admissions to protect access for Medicaid beneficiaries. When increasing Page 6 December 29, 2017

11 Arkansas DRG Conversion Plan payment for types of services, policy adjustors are used. There are four types of adjustors commonly used, and should be considered as options: Service adjustors Age/service adjustors Provider/service adjustors Provider adjustors If implementing all four options for policy adjustors, the calculation of DRG base payment becomes: [DRG Base ] = [Hospital Base Rate] * [DRG Relative Weight] * [Service Adjustor] * [Age/Service Adjustor] * [Provider/Service Adjustor] or [DRG Base ] = [Hospital Base Rate] * [DRG Relative Weight] * Maximum of ([Service Adjustor], [Age/Service Adjustor], [Provider/Service Adjustor]) Policy adjustors, in general, modify payment for specific types of services, patient ages and hospital types. Service adjustors apply for specific types of care independent of the recipient and provider. Age/service adjustors apply only for recipients within a specific age range. Any age range can be used, but Medicaid programs generally use this to increase payment for pediatric care. For Arkansas Medicaid, an age adjustor for children under the age of 1 would provide higher payment for these recipients consistent with current policy. Provider/service adjustors apply for specific services and only when care is delivered at a certain category of hospitals. For example, if a Medicaid agency decided to increase payments for neonatal care using a service adjustor of 1.5, then the claim payment would be increased by 50 percent. In this situation, a claim submitted from a provider with base rate $7,000 and mapping to APR DRG (Neonate birth weight grams with major respiratory condition; relative weight = in version 35) the DRG base payment would be calculated as follows: [Maximum Adjustor] = Max(1.5, 1.0, 1.0) = 1.5 [DRG Base ] = $7,000 * * 1.5 = $33, As a separate example, a Medicaid agency might decide to increase payment for pediatric care using an age/service adjustor of In that case, a claim submitted from a provider with base rate $7,000, for a recipient age 10, and mapping to APR DRG (Asthma; relative weight = in version 35) the DRG base payment would be: [Maximum Adjustor] = Max(1.0, 1.25, 1.0) = 1.25 [DRG Base ] = $7,000 * * 1.25 = $4, Page 7 December 29, 2017

12 Arkansas DRG Conversion Plan A separate claim from the same hospital for a recipient age 35 (above the age adjustor cut-off) and mapping to the same APR DRG, 141-2, would generate a DRG base payment of: [Maximum Adjustor] = Max(1.0, 1.0, 1.0) = 1.0 [DRG Base ] = $7,000 * * 1.0 = $3, Adjustments to DRG Base Transfer Claims When processing claims for recipients transferred from one acute facility to another, most Medicaid DRG implementations have followed the Medicare model for payment adjustments. In this model, a payment amount is calculated using a per diem method and then compared to the DRG base payment. If the per diem payment, referred to as a transfer-adjusted base payment, is less than the DRG base payment, then the transfer-adjusted base payment is used. Using the DRG base payment and the DRG s average length of stay, a transfer-adjusted payment can be calculated as: Transfer-Adjusted Base = {[DRG Base ] / [DRG Average Length of Stay]} * {[Length of Stay] + 1} Adding one to the length of stay takes into account the disproportionate amount of costs required in the first day of admission to complete the admission process and perform an initial diagnostic evaluation. For example, APR DRG (neonate birth weight grams with respiratory distress syndrome, other major respiratory anomaly or other major anomaly) has relative weight and average length of stay equal to days (in version 35). If a baby with this DRG is transferred out of a hospital after two days and the hospital s base price is $7,000 then, Full DRG Base = $7,000 * = $66, Transfer-Adjusted Base = (66, / 57.88) * (2 + 1) = $3, In this example, the transfer-adjusted base payment is less and would be used in place of the full DRG base payment Partial Eligibility If a recipient is only eligible for Medicaid payment for part of a hospital stay, then a full DRG payment may not be appropriate. A smaller payment may be acceptable as the hospital will be getting reimbursement for part of the stay from other sources, such as a Medicare. in a partial eligibility situation can be determined using the same approach as transfer claims a per diem payment is calculated, compared to the full DRG base payment, and the lower of the two is used. The calculation of eligibility-adjusted base payment can be exactly the same as the transfer-adjusted base payment. That is, Eligibility-Adjusted Base = {[DRG Base ] / [DRG Average Length of Stay]} * {[Length of Stay] + 1} Page 8 December 29, 2017

13 Arkansas DRG Conversion Plan Another option is to remove the + 1 from the number of days multiplier in cases where the Medicaid eligibility did not begin until after the day of admission. In that case the formula is, Eligibility-Adjusted Base = {[DRG Base ] / [DRG Average Length of Stay]} * [Length of Stay] in a partial eligibility situation can also be determined using a different method a proration based on the number of days for which the recipient had eligibility. Under this method, a simple percentage is calculated by dividing the number of days of eligibility by the total days of the hospital stay. And then the full DRG payment gets reduced by this percentage. The formula under this method is, Eligibility-Adjusted Base = {[Medicaid Covered Days] / [Length of Stay]} * [DRG Base ] Another possible proration formula compares Medicaid Covered Days to the DRG s Average Length of Stay, and pays less than full DRG payment only if the number of Medicaid Covered Days is Less than the DRG Average Length of Stay as follows: Eligibility-Adjusted Base = Minimum of {[Medicaid Covered Days] / [DRG Avg Length of Stay]} * [DRG Base ] And [DRG Base ] Partial eligibility scenarios may occur for a variety of reasons, including, Medicare Part A benefit expires in the middle of a hospital inpatient stay for a recipient dually eligible for Medicare and Medicaid A recipient fulfills spend down requirements during the middle of a hospital inpatient stay and thus, becomes eligible for Medicaid coverage for the later portion of a hospital admission Recipient is an undocumented alien, in which case Medicaid is only responsible for reimbursement for the emergency portion of the hospital stay 3.6 Provider Loss Outlier s Inevitably, some claims will be submitted for extreme and unpredictable cases in which the standard DRG payment differs greatly from the level of resources expended by the hospital. For these cases, referred to as outliers, a DRG payment method can adjust payment upward to share in hospital losses. The Medicare model, also adopted by several states, is a cost-based outlier policy that employs a stop-loss threshold which generates outlier payments whenever the hospital s estimated loss is above a threshold. With this method, the formula for an outlier payment adjustment is: [Hospital Loss] = ([Billed Charges] * [Cost to Charge Ratio]) - [DRG Base ] If [Hospital Loss] > [Outlier Threshold] Then [Outlier Pymt Adjstmnt] = ([Hospital Loss] [Outlier Threshold]) * [Marginal Cost %] Else [Outlier Adjstmnt] = 0 Page 9 December 29, 2017

14 Arkansas DRG Conversion Plan For example, an admission with charges of $200,000, at a hospital with cost-to-charge ratio equal to 0.30 and a DRG base payment of $5,000 has a hospital loss equal to $55,000 {($200,000 * 0.3) - $5,000}. If the Medicaid DRG policy included an outlier threshold of $30,000 and a marginal cost percentage of 70 percent then the outlier payment would be {($55,000 - $30,000) * 0.7) = $17,500. Thus the final payment to the provider would be ($5,000 + $17,500) = $22,500. Some states also implement length of stay outlier payments. For example, Mississippi Medicaid uses a length of stay outlier policy for mental health services. If a hospital admission assigned a mental health APR DRG has a length of stay greater than 19 days, the Mississippi Medicaid policy applies an outlier per diem payment to days 20 through date of discharge. 3.7 DRG Price versus Final Reimbursement The previous sections in Chapter 3 describe how the DRG price is calculated. This is the amount of money Medicaid is willing to pay for the services without consideration of any other forms of payment. This price is sometimes referred to as the Medicaid allowed amount. Final reimbursement for a claim equals the DRG price minus any other forms of payment such as payment from another insurance carrier, recipient spend down, and patient cost sharing, such as copays. Thus, [Final Reimbursement] = [Allowed Amount] [Other Ins Pymt] [Spend Down] [Cost Sharing] 3.8 Non-DRG Paid Claims Depending on the payment policies set by the state, some acute care inpatient claims may fall outside the DRG payment. These may be claims for services or providers carved out of the DRG payment method, or they may be interim claims from providers for services that are included in DRG payment. Both carved out items and interim claims are commonly paid using a per diem model, although they can also be paid as a percentage of charges. If a per diem is used to reimburse interim claims, the per diem is set relatively low as it is intended to be a temporary, partial payment. The interim claim per diem gives hospitals some reimbursement for cash flow purposes, while still leaving the hospital incentive to submit a final claim when the recipient is discharged for final DRG payment. Page 10 December 29, 2017

15 Arkansas DRG Conversion Plan 4 Scope of DRG Method 4.1 Affected Providers Affected Providers - Discussion DRG payment methods typically cover payments to general acute care inpatient facilities. Nursing home care and hospice care are normally paid outside of a DRG payment method. There are other provider types, however, where the decision of inclusion or exclusion in DRG payment is less clear and varies among states using DRG payments. These provider types include: Physical rehabilitation Long term acute care Mental health and substance abuse facilities Psychiatric residential treatment facilities Critical access or rural hospitals Children s hospitals Cancer hospitals In-state / out-of-state / border hospitals Native American Indian hospitals Public hospitals The first three provider types in the list above, physical rehabilitation, long term acute care, and mental health / substance abuse facilities all treat patients with highly variable and unpredictable lengths-of-stay. Because of this, some states choose to pay these providers with another method, such as a per diem method, instead of paying via DRGs. In addition, a hybrid option is possible where providers are paid per diem and the per diem amount is adjusted based on patient acuity, using DRG grouping to measure patient acuity. The APR DRG patient classification model, for example, contains 72 different APR DRG classifications and relative weights intended to reflect the resource intensity of different types of psychiatric patient care. The relative weights associated with the APR DRG classifications can be used to adjust the per diem, offering a higher per diem for above average relative weight and a lower per diem for below average relative weight. The next three providers, critical access, children s, and cancer hospitals are all excluded from the Medicare DRG inpatient prospective payment system. For that reason, states can encounter resistance when including these providers in the Medicaid DRG payment method. simulations are a valuable tool for reviewing payments to these providers under a DRG method and help to show whether or not DRGs will offer fair reimbursement. With the robustness of some DRG models, such as that reflected in the APR DRG algorithm, the simulations often do show DRG payment is a reasonable option. In addition, special considerations within the DRG payment method can be reviewed to ensure fair reimbursement if needed. For example, separate hospital base rates can be given for some or all of these categories of providers. Also, certain services can be given a service or age adjustor. In addition, certain services can be defined as separately billable on outpatient claims, such as organ search and acquisition costs, and blood factors, which is particularly appealing to cancer institutions. Making these kinds of payment adjustments within the overall DRG payment Page 11 December 29, 2017

16 Arkansas DRG Conversion Plan method allows for special considerations to be made while still maintaining the simplicity of all or nearly all providers being paid using the same method. Similarly, to maintain simplicity, most states pay in-state, border hospitals, and out-of-state hospitals via DRGs. The only decisions normally made based on general location of each hospital are selection of hospital base price and determination of cost-to-charge ratio. For outof-state hospitals, normally a single hospital base price and a default cost-to-charge ratio are used. For example, the state s standard Medicare urban or rural cost-to-charge ratio can be assigned to each out-of-state hospital. However, border hospitals may have a sufficiently high volume of Medicaid recipients to justify treating them like in-state hospitals for the purpose of assigning base rates and cost-to-charge ratios. Finally, many Medicaid agencies have separate policies associated with Native American Indian hospitals and public hospitals, so decisions need to be made on how these categories of providers will be affected by a DRG payment method. 4.2 Affected Services Affected Services - Discussion The list of services sometimes included and sometimes excluded from DRG payments is similar to the list of provider types open for debate. States vary on inclusion in DRG payment for the following list of services: Physical rehabilitation Mental health and substance abuse Unpredictable and expensive services and supplies such as blood factors and organ search and acquisition New technologies As described in the previous section, a policy decision must be made relating to inclusion or exclusion of specialty rehabilitation and psychiatric institutions within a DRG payment method. In addition, a policy decision must be made for payment of rehabilitation and psychiatric services when performed within general acute care facilities. If volumes are low, the simplicity of including them in the DRG payment method are likely justifiable. However, if volumes are high, it will be more justifiable to pay these services using the same methodology as the specialty institutions and/or distinct part units (which may be DRG-exempt). Unpredictable and expensive services and supplies such as blood factors and transplant organ searches create challenges for a DRG payment method. DRG payments are based on average resource usage and work very well when hospital admissions can be grouped into relatively homogeneous categories. However, some cases require resources far outside the norm, such as the cost of blood factors required when operating on a patient with a blood clotting problem. Medicare as an example has taken the stance that some unpredictable and/or expensive services do warrant payment above and beyond DRG payment. Specifically, Medicare s inpatient prospective payment system allows for separate payment for inpatient services under three circumstances: Organ acquisition. In most cases, these costs are reimbursed through the cost settlement process; for renal transplants, designated renal transplantation hospitals are paid adjusted rates. Page 12 December 29, 2017

17 Arkansas DRG Conversion Plan Blood clotting factors. Blood factors are paid based on a fee schedule (e.g., 95 percent of average wholesale price). New medical technology. Devices that meet very specific Medicare criteria related to newness, FDA approval, substantial clinical improvement and unusual costliness criteria, may qualify for add-on payments. Very few devices meet these criteria. State Medicaid DRG payers, in contrast, often do not allow separate payment for unpredictable and expensive services because of both the concern over incentives and the added complexity to the payment method. From the point of view of Arkansas Medicaid, items that occur in very low volumes might be reimbursed sufficiently through outlier payments. However, if volumes are high or are heavily concentrated at specific hospitals, outlier payments alone may not be sufficient. Instead, certain services and supplies can be carved out of the DRG payment and made separately payable. However, such a policy can be extremely challenging to implement in an MMIS. Other options such as different provider base rates, service adjustors, or multiple tiers in the outlier payment method (using a higher marginal cost percentage for very high losses) may generate fair payment and prove far simpler to implement. New technologies can also be a challenge for a DRG payment. In theory, they may reduce cost of care, but in practice, they most often increase cost. Furthermore, DRG relative weights may lag slightly behind in capturing these costs because DRG relative weights are calculated using costs from historical claims. Thus, offering separate payment for new technologies is justifiable. However, the task of maintaining an ever-evolving list of new technologies is very challenging. In addition, estimating the budgetary impact to Medicaid for separately reimbursing new technologies is difficult to do accurately. As with many policy decisions, the topic of unpredictable and expensive services requires a trade-off between the principles used to evaluate a payment method (described in Chapter 1). Allowing separate payment for unpredictable and expensive services diminishes the incentives for efficiency, reduces transparency, increases administrative burden, and increases complexity. On the other hand, access to care may be jeopardized if certain types of cases result in predictable and consistent losses, even with the case mix and outlier adjustments of a DRG payment method. An example is surgery for patients with hemophilia. The need for blood factors can sharply increase the hospital s cost even for otherwise routine surgeries. 4.3 Affected Beneficiaries / Medicaid Programs Affected Beneficiaries / Medicaid Programs - Discussion Medicaid agencies generally administer a variety of programs, usually with beneficiaries enrolled in only one program at a time. Common programs include fee-for-service, primary care case management, managed care, and Children s Health Insurance Program (CHIP). States often also administer smaller programs sometimes based on a waiver and sometimes paid for by separate funding sources than used for standard Medicaid. In addition, some Medicaid beneficiaries are eligible only for specific services, most notably emergency-only services. Lastly, some Medicaid beneficiaries are dually eligible for Medicaid and Medicare. For these beneficiaries, most healthcare services are paid primarily by Medicare with Medicaid acting as a supplementary payer, usually paying only the Medicare coinsurance and deductible amounts. However, there are certain services not covered by Medicare and cases where Medicare Page 13 December 29, 2017

18 Arkansas DRG Conversion Plan benefits have been exhausted, in which case Medicaid becomes the primary payer. As part of a DRG payment method implementation, Medicaid agencies must determine which programs and/or eligibility categories will be included in the new payment method. The new payment policy must also decide how Medicare crossover claims (where Medicare was the primary payer) are affected. For simplicity of the payment methods, Medicaid programs typically aim to include all programs in the DRG payment method and make exceptions only when specific, justifiable reasons are identified. 4.4 Billing and Review Changes Billing and Review Changes - Discussion When moving from a per diem-based payment method to a DRG-based payment method, the unit of service that is tied to the payment methodology changes from a day to a complete hospital stay, or discharge. This often has implications on the service authorization process. In a per diem payment method, processes and systems are often installed to monitor the number of days of each hospital stay. Under a DRG payment method, length of stay is no longer a major contributor to payment. As a result, the Medicaid program no longer needs to emphasize careful control over the number of days authorized. Instead, Medicaid programs using a DRG payment method generally choose only to authorize hospital admissions, not the number of days following the admission. Thus, pre-admission review processes often change when shifting from per diem to DRG payment. Similarly, post-admission review processes often change when shifting from per diem to DRG payment. Under a DRG payment method, hospital incentives to maximize reimbursement are different, and payment reviews should be adjusted accordingly. Medicaid programs using DRGs benefit from monitoring volumes of very expensive stays to avoid excessive outlier payments. Medicaid agencies using DRGs also monitor volume of unusually short stays to prevent inappropriately early discharges, and monitor volume of readmissions as hospitals receive separate payment for each admission/discharge occurrence. DHS will also need to consider developing a readmission policy under the DRG payment methodology, as DRG-paid readmission payments are more impactful than short-stay per diempaid readmissions under the current methodology. There are a number of types of readmission policies for consideration, including a) Claim Denial- or Consolidation-Based Policies, or b) Performance Measurement with Prospective Adjustment. The first policy type focuses on payment penalties for individual readmissions deemed to be related to the initial admission, whereas the second policy type is a wholistic population-based approach focusing on rates of readmissions over time and provider performance relative to benchmarks. policies that create incentives for reducing avoidable readmissions can deliver substantial cost savings, while at the same time providing the administrative capacity to measure and regulate the quality of care delivered to highly-vulnerable patients in acute and post-acute care (PAC) settings. From the providers perspective, hospitals will need to report diagnosis codes completely and in the correct order to receive appropriate payment under APR DRGs. For example, in the raw MMIS inpatient claims data we observed several thousand delivery-related claims where the reported primary diagnosis code was the delivery result (e.g. single live born) rather than the reason for admission (e.g. encounter for full-term delivery). Without adjustments, these claims would not receive a delivery-related DRG as appropriate (we made adjustments in the DRG model to appropriate assign these DRGs). Under actual claim payment processing, the MMIS will not make such adjustments; as such we expect providers will respond to a new DRG Page 14 December 29, 2017

19 Arkansas DRG Conversion Plan system by improving their capture of the patient medical record and report diagnosis codes comprehensively and in the correct order. Page 15 December 29, 2017

20 Arkansas DRG Conversion Plan 5 DRG Grouping The topic of DRG grouping breaks down into two major decision points. The first is which DRG grouping algorithm to use. Once that is decided, then the source of the DRG relative weights and average lengths of stay can be determined. 5.1 DRG Grouper DRG Grouper - Discussion Introduction The goal of diagnosis related groupers is to define patients into categories based on similar clinical conditions and on similar levels of hospital resources required for treatment. These categories are identified using DRG codes, each of which is assigned a relative weight appropriate for the relative amount of hospital resources used to treat the patient. For example, if a DRG grouper assigns patient A to DRG 123 with relative weight 0.5, and assigns patient B to DRG 321 with relative weight 1.0, this indicates the average amount of hospital resources required to treat patient A is a half the amount of resources required to treat patient B. These relative weights associated with DRGs are used in the calculation of reimbursement with the intent of paying more when the patient s care required more resources and less when the patient s care required fewer resources. Thus, from the point of view of hospital reimbursement, the best DRG grouper for a particular healthcare payer is the one that most accurately predicts the relative hospital resource usage for the full range of services reimbursed by the payer. Given the importance of generating fair payment for services provided, the primary objective of a DRG grouper is to categorize hospital stays in a way that most accurately predicts relative hospital resource usage for the care provided to each patient. In addition, there are other benefits of DRG grouping such as contributing to measurement of hospital quality and categorizing the types of care reimbursed by the payer. Also, as with any tool, DRG groupers need to be evaluated in terms of long term viability and reliability. With all these thoughts in mind, the criteria recommended for evaluation of different DRG groupers are: 1. Accuracy categorizing relative cost of care for the full range of services reimbursed by the Medicaid agency, with particular concentration on the services for which Medicaid is a major player in the market 2. Long term viability in an ever-evolving healthcare industry 3. Ability to contribute to measurement of hospital quality 4. Familiarity and experience being used in the industry Options Given the need to for long-term viability and experience being used in the industry, we believe there are currently only two DRG algorithms available that are worthy of consideration. Those are All Patient Refined Diagnosis Related Groups (APR DRGs) and Medicare Severity Diagnosis Related Groups (MS-DRGs). 1 1 Other DRG algorithms include CMS-DRGs, AP-DRGs, Tricare DRGs, and APS-DRGs. CMS-DRGs and AP-DRGs have been phased out. Neither is actively being updated and neither was released with an ICD-10 compliant version. The Tricare DRG algorithm, which was developed and is currently maintained by 3M, uses generally the same DRG grouping logic as MS-DRGs, but has been enhanced to reflect the grouping logic of the obsolete AP-DRG model for pediatric and neonatal services. Based on our discussions with representatives from 3M, there has been relatively Page 16 December 29, 2017

21 Arkansas DRG Conversion Plan These are compared in greater detail in Table 5.1. Table 5.1 Detailed Comparison of Select DRG Algorithms Description MS-DRGs V.35 (CMS - Maintained by 3M) APR DRGs V.35 (Maintained by 3M and NACHRI) Intended population Medicare (age 65+ or under age 65 with disability) All patient (based on the Nationwide Inpatient Sample) Overall approach and treatment of complications and comorbidities (CCs) Intended for use in Medicare Population. Includes 338 base DRGs, initially separated by severity into no CC, with CC or with major CC. Low volume DRGs were then combined. Structure unrelated to Medicare. Includes 326 base DRGs, each with four severity levels. The is no CC or major CC list; instead, severity depends on the number and interaction of CCs. Number of DRGs 752 valid codes 1,304 valid codes Newborn DRGs 7 DRGs, no use of birth weight 28 base DRGs differentiated by birth weight range and other conditions, each with four levels of severity (total 112) Psychiatric DRGs (including chemical dependency) 11 DRGs 18 DRGs, each with four levels of severity (total 72) use by Medicaid KS, NC, NH, NM, OK, OR, SD, WV Operational: AZ, CA, CO, CT, DC, FL, KY, IL, IN, MI, ND, NE, OH, MA, MD, MS, MT, NY, PA, RI, SC, TX, VA, WI, WA Announced or Considering: AL, WY Other users Medicare, hospitals Hospitals, AHRQ, MedPAC, JCAHO, various state report cards Uses in measuring hospital quality Used as a risk adjustor in measuring readmissions. Used to reduce payment for hospital-acquired conditions. Used as risk adjustor in measuring mortality, readmissions, complications little investment focused on the Tricare DRG tool to bring it current with the standards established for more current models, particularly with respect to classifying neonatal and pediatric cases. The DRGs for those types of cases have been the same for many years and have not been (nor are they expected to be) updated with new research. Finally, we are not aware of any payer in the United States using APS-DRGs for payment purposes. For these reasons, we consider the CMS-DRG, AP-DRG, Tricare DRG, and APS-DRG algorithms to be unacceptable options. Page 17 December 29, 2017

22 Arkansas DRG Conversion Plan Accuracy Categorizing Relative Cost with a Medicaid Population The APR DRG algorithm is designed for a full beneficiary population. In addition, it includes significant granularities for sick newborns and pediatrics that are developed and maintained by the National Association of Children s Hospitals and Related Institutions (NACHRI) for 3M Health Information Systems. This is a good fit for a Medicaid population, which is heavily weighted towards newborns and children in the traditional model (excluding the Affordable Care Act (ACA) expansion population). MS-DRGs, in contrast, are developed specifically for the elderly Medicare population. The DRGs are designed for beneficiaries over the age of 65 or who are disabled or suffering from end stage renal disease. In 2004 the Centers for Medicare and Medicaid Services (CMS) made a policy shift to no longer support the needs of all payers: As previously stated, we do not have the data or the expertise to develop more extensive newborn and pediatric DRGs. Our mission in maintaining the Medicare DRGs is to serve the Medicare population. 2 In 2007 CMS adopted its new Medicare Severity DRG algorithm (MS-DRGs) and made several statements underscoring the fact that MS-DRGs were developed only for the elderly Medicare population. Specifically, CMS noted: The MS-DRGs were specifically designed for purposes of Medicare hospital inpatient services payment. As we stated above, we generally use MEDPAR data to evaluate possible DRG classification changes and recalibrate the DRG weights. The MEDPAR data only represent hospital inpatient utilization by Medicare beneficiaries. We do not have comprehensive data from non-medicare payers to use for this purpose. The Medicare program only provides health insurance benefits for people over the age of 65 or who are disabled or suffering from end-stage renal disease. Therefore, newborns, maternity, and pediatric patients are not well represented in the MEDPAR data that we used in the design of the MS-DRGs. We simply do not have enough data to establish stable and reliable DRGs and relative weights to address the needs of non-medicare payers for pediatric, newborn, and maternity patients. For this reason, we encourage those who want to use MS-DRGs for patient populations other than Medicare make the relevant refinements to our system so it better serves the needs of those patients. 3 The number of newborn DRGs provides a useful contrast between the MS-DRG algorithm and an all-patient algorithm such as APR DRGs. MS-DRGs provide seven (7) DRG codes for the care of newborns while APR DRGs provide 112 DRG codes (28 base DRGs, each with four (4) levels of severity). In addition, MS-DRGs do not take birth weight into consideration when assigning a DRG despite the fact that birth weight has been widely accepted as a significant indicator of the viability and overall health of newborns (whereas APR DRGs use birth weight when assignment newborn DRGs). 2 CMS, Medicare Program; Changes to the Hospital Inpatient Prospective Systems and Fiscal Year 2005 Rates; Final Rule, Federal Register 69:154 (Aug. 11, 2004), p. 48, CMS, Medicare Program; Changes to the Hospital Inpatient Prospective Systems and Fiscal Year 2008 Rates; Final Rule, Federal Register 72:162 (Aug. 22, 2007), p. 47,158. Page 18 December 29, 2017

23 Arkansas DRG Conversion Plan Long Term Viability As mentioned previously, CMS-DRGs and AP-DRGs have already been discontinued and are not compatible under ICD-10 coding. APR DRGs and MS-DRGs are heavily used, and widely accepted, so their viability is strong. Both have been released with ICD-10 compliant versions and are expected to be updated as necessary to follow future changes in healthcare payment strategies in the United States for years to come Applicability to Quality Measures Incorporating hospital quality measures into payment systems has become increasingly common and sophisticated over the past decade. States face increasing pressure to demonstrate that Medicaid payments support quality care. 4 To fairly measure hospital quality, the quality measure should be risk adjusted (also referred to as case mix adjusted). For example, performing direct comparisons of mortality rates or complication rates between a cancer institute and a small rural hospital would be unfair unless they are case mix adjusted. In a situation where a cancer institute has a complication rate of 7 percent, and a small rural hospital has a complication rate of 5 percent, at face value, the complication rate of the cancer institute appears higher. However, when taking into consideration patient acuity between the two facilities, the complication rate at the cancer institute might prove to be lower than the rate at the rural hospital. APR DRGs are very commonly used for the purpose of case mix adjustment. APR DRGs are also used as a basis for two quality measurement tools becoming increasing popular with Medicaid programs for measurement of hospital quality using medical claims data. Those tools are: 3M Potentially Preventable Complications (PPC) Grouping Software identifies complications that may have been avoided. This software first identifies conditions not present on admission and then determines whether those conditions were potentially preventable given the patient s reason for admission, procedures, and underlying medical conditions. It also flags Hospital Acquired Conditions monitored by CMS. 3M Potentially Preventable Readmission (PPR) Grouping Software identifies readmissions clinically related to previous admissions which were potentially preventable. Both of the above software applications are currently used by several Medicaid agencies as part of payment policies that create incentives for reducing avoidable readmissions and associated costs, while at the same time providing the administrative capacity to measure and regulate the quality of care delivered to highly-vulnerable patients in acute and post-acute care (PAC) settings. Because the 3M PPC and PPR quality measurements are built using the language of APR DRGs, implementing APR DRGs for payment can facilitate a move to PPC and PPR quality measures. 4 Evidenced by section 2702 of the Patient Protection and Affordable Care Act prohibiting federal Medicaid payments for services treating healthcare-acquired conditions (effective July 1, 2012). Page 19 December 29, 2017

24 Arkansas DRG Conversion Plan Prevalence in the Industry APR DRGs are by far the preferred choice for state Medicaid agencies because of their applicability to the Medicaid population. Figure 5.1 shows how states currently pay for inpatient care, and shows over 25 state agencies currently using APR DRGs. APR DRGs have also been used to adjust for case mix differences in performance measures in Arkansas, Hawaii, Maryland, Massachusetts, New York, Texas and Utah. 5 MS-DRGs are the DRG algorithm implemented for Medicare. In addition, twelve state Medicaid agencies have chosen MS-DRGs and some, if not all of these, add customized state-only DRGs or other tools to more accurately categorize care for newborns. Figure 5.1: Medicaid Inpatient Methodologies by State DRG Grouper - Recommendation For a Medicaid population, the granularity and focus on newborn, maternity and pediatric services in the APR DRG grouper makes it the best option for use in inpatient claim reimbursement. The only other prevalent option in the industry, MS-DRGs are not well suited for a Medicaid population (at CMS s own admission). APR DRGs, in contrast, have sufficient granularity to categorize hospital stays and associated cost of care for the full range of beneficiaries served by Medicaid agencies. In fact, APR DRGs are particularly detailed for certain services in which Medicaid is a major payer, specifically sick newborns (neonates), 5 Prepared by ACS for the California Department of Health Care Services. Medi-Cal DRG Project Draft Policy Design Document. January 10, Page 24. Page 20 December 29, 2017

25 Arkansas DRG Conversion Plan obstetrics and pediatrics. APR DRGs are currently used by numerous state Medicaid agencies for claims payment and are planned for implementation in a few additional states. With its strong market share, APR DRGs are expected to be updated for future changes impacting the U.S. medical insurance industry, as was done with the implementation of ICD-10. And finally, APR DRGs are heavily used for risk adjustment and for hospital quality measures that are becoming more prevalent as a way to incent quality care. 5.2 DRG Relative Weights DRG Relative Weights - Discussion States have three options when selecting a set of relative weights for the DRGs they will be using: a. Use national relative weights b. Develop state-specific weights c. Borrow state-specific weights developed by another payer or Medicaid program National relative weights exist for APR DRGs, and MS-DRGs. For APR DRGs, 3M publishes national relative weights annually for each APR DRG grouper version release. 3M calculates national weights using the two most recent year s data from the National Inpatient Sample (NIS) maintained by the Agency for Healthcare Research and Quality (AHRQ). This data includes claims from all payer types (including Medicaid) and also includes Arkansas hospital experience. 3M calculates two sets of national weights for each APR DRG grouper version release, as follows: Standard weights: Calculated based on the average charges per discharge for each APR DRG divided by the national average charge per discharge (for all APR DRGs). The standard approach is the simplest and most commonly used of the two sets of national weights; however, it does not consider charge basis variation across facilities. Hospital-Specific Relative Value, or HSRV weights: Alternative APR DRG weight set that measures and adjusts for the charge basis variation across hospitals. Like standard weights, the HSRV weights are based on average charges per DRG, except the charges are adjusted based on a regression analysis to standardize the charge basis across all facilities in the underlying NIS data. Due to its complexity, HSRV weights have not been adopted by State Medicaid agencies like the standard weights. MS-DRG relative weights are also published each year by CMS for each MS-DRG grouper version release. CMS determines its MS-DRG weights using claims data from Medicare recipients (MEDPAR data). National weights are relatively easy to adopt as they are calculated by external agencies. If using national relative weights, states can decide to use the raw values as they are published, or re-center the weights to the individual state s overall case mix. Re-centering the weights simply resets the average relative weight to 1.0 which makes the numbers easy to understand relative weights less than 1.0 are below average and relative weights above 1.0 are above average. However, re-centering of the weights adds a small amount of administrative Page 21 December 29, 2017

26 Arkansas DRG Conversion Plan complexity and makes comparison of Arkansas Medicaid case mix measurements to national benchmarks more difficult. Instead of using national relative weights, states can choose to calculate their own weights. This option has the benefit of ensuring the weights accurately reflect costs of hospitals when treating patients that are unique to that state s Medicaid population. However, calculating statespecific weights requires significantly more effort from the Medicaid agency each time the State adopts a new grouper version. In addition, it offers the challenge of deciding what values to use for DRGs with statistically low volume in the Medicaid program. Even California, the largest Medicaid program in the country, found there were 463 APR DRGs with fewer than 30 stays in a single year (2009), including 46 APR DRGs with zero volume. 6 Arkansas Medicaid has significantly less enrollment than California and would have significantly more APR DRGs with insufficient volume to be used in calculation of state-specific weights. In cases with low volume, states can choose to use the national value, or prorate the weight from a similar DRG. If choosing to use state-specific relative weights, decisions must also be made on how those weights will be calculated. The basis for weights can be charges or relative costs. Typically, relative weights come out similarly when using charges or costs, but using costs is far more defensible due to wide variation in hospital charge levels. When using costs, another necessary decision is defining how costs will be determined for the relative weight calculation. Further, the process for recalculating the weights would have to be performed periodically, usually annually. The final option a state can select is to copy the relative weights from another Medicaid program. This has the advantage of limiting the effort a state expends to determine relative weights while allowing the weights used to be specific to a Medicaid program. Pennsylvania selected this option, and uses the state-specific APR DRG relative weights calculated by New York. Once a DRG grouper is selected, a comparison can be made of national relative weights versus state-specific weights. Navigant has performed this type of comparison in the past and found the national weights and state-specific weights align very closely on the high volume and high cost DRGs. Similar to relative weights, average length of stay must also be determined for each DRG. Average length of stay is used in transfer and partial eligibility payment adjustments. Average length of stay can also be used in outlier calculations if day outliers are implemented. If using national relative weights, national average lengths of stay would also be available for use. Similarly, if borrowing from another state, both the relative weights and average lengths of stay could be borrowed. If, on the other hand, Arkansas Medicaid state-specific relative weights are selected, then state-specific average lengths of stay would also need to be calculated, including the challenge of deciding what to do with DRGs having statistically low volumes of observations DRG Relative Weights - Recommendation Studies with other state Medicaid data have shown that state-specific weights and national weights align very well for high volume DRGs. An example of this type of analysis is shown in Figure 5.2 using Florida-specific relative weights. Also, as mentioned in the discussion section, using national weights requires less administrative burden and requires little or no manual adjustment for low-volume DRGs. 6 Prepared by ACS for the California Department of Health Care Services. Medi-Cal DRG Project Draft Policy Design Document. January 10, Page 33. Page 22 December 29, 2017

27 National APR-DRG Rel Wts Re-centered to FL Medicaid Stays Arkansas DRG Conversion Plan Figure 5.2: Comparison of re-centered APR DRG standard national weights versus Florida Medicaid relative weights APR-DRG Relative Weight Comparison Top 50 Florida Medicaid APR-DRGs By Claim Volume R 2 for top 50: R 2 for all DRGs: (Stays >= 20) Top 50 DRGs account for 6 of all stays Florida Medicaid APR-DRG Relative Weights APR-DRG with FL weight 4.63 and national weight 4.13 is not shown We have observed strong correlation in similar analyses in Arizona, California, Illinois, Minnesota, Washington and Wisconsin. Given the strong correlation between Medicaid-specific weights and 3M national weights and the sample size issues from creating state-specific weights, we recommend adopting 3M s national weights for use by the Arkansas Medicaid program. Given the simplicity and widespread adoption of the standard weight set by Medicaid agencies compared to the HSRV weights, we recommend the standard national weight set for the Arkansas Medicaid program. In addition, we do not recommend re-centering the national weights to 1.0 (by dividing each national relative weight by the Arkansas Medicaid overall average case mix). Re-centering the weights offers only a minor benefit of making 1.0 the average Arkansas Medicaid relative weight, numbers below 1.0 less than the average, and numbers above 1.0 greater than the average. At the same time, re-centering makes Arkansas weights different than national and other state weights making comparisons across states difficult. In addition, use of re-centered weights increases complexity because it requires all interested parties, including large hospital Page 23 December 29, 2017

28 Arkansas DRG Conversion Plan chains and managed care plans who may already use APR DRGs, to use a separate, customized set of weights specifically for Arkansas Medicaid. If Arkansas state-specific weights are used, there are 670 APR DRGs with volume below 20 stays in one year of fee-for-service data (using data from calendar year 2016). States typically use a minimum threshold between 10 and 30 stays as the minimum volume needed to calculate a state-specific relative weight. For this discussion, we have picked the midpoint of this range, or 20 stays. DRGs with less than 20 stays would need their relative weights determined using data borrowed from another source as a proxy. Also, any shift to managed care would result in the number of DRGs in the fee-for-service population that have a sample size of less than 20 stays to increase. Given these issues, state-specific weights would likely need to be calculated in the future using a combination of Medicaid fee-for-service and managed care claims to reach sufficient volume for relative weight calculations. Page 24 December 29, 2017

29 Arkansas DRG Conversion Plan 6 Provider Base Rates Provider base rates are another significant contributor to the reimbursement amount on individual hospital stays and to Medicaid hospital inpatient reimbursement in aggregate. Thus selection of provider base rates is a critical step in ensuring fair reimbursement when implementing a DRG payment method. The simplest approach from the point of view of maintaining budget neutrality would be to assign each hospital its own base rate. However, this would defeat one of the basic goals of a DRG payment method that is incentivizing and rewarding hospital efficiency. The opposite approach would be to develop a single base rate to be applied to all hospitals. Many Medicaid agencies have found a more practical approach is something in the middle in which a small number of standardized DRG base rates are created for a small number of categories of hospitals to address reasonable and measurable differences in cost. Some states further adjust for hospital cost by applying hospital-specific wage indices to the standardized base rates. 6.1 Provider Base Rate Categories Provider Base Rate Categories - Discussion Many Medicaid agencies implementing a DRG payment method have chosen to select different DRG base rates for different hospital categories or peer groups. Base rates determined separately for hospital peer groups can aid in protecting access to care at specific facilities, such as critical access, rural hospitals, and children s specialty hospitals. In addition, when looking at cost structures, separate base rates may be justifiable, for example, for trauma facilities and/or teaching hospitals. For teaching hospitals, Medicare provides additional payment, separate from the base rate. However, that additional payment can just as easily be incorporated into the base rate. 7 A peer group can also be considered if there is a group of hospitals who treat very complicated, expensive cases and are expected to have an unusually high percentage of outlier payments. In most DRG implementations, outlier payments cover a lower percentage of hospital costs than standard DRG payments so high numbers of outlier stays become a burden to hospitals. One way to solve that problem is to give these hospitals a higher base rate, which will serve to reduce their percentage of outlier stays. If separate base rates are selected for some groups of providers, we recommend the criteria used to categorize hospitals within groups be very clear and maintainable. Understandably, hospitals will be motivated to be defined into the peer group offering the most attractive reimbursement. Having clearly defined criteria for each grouping will help maintain the integrity of the payment policy and lessen the administrative burden of categorizing hospitals. 7 For example, the California Department of Health Care Services implemented a separate base rate for remote rural hospitals to protect access to care. Florida Medicaid applied multipliers that increase the DRG base rates for rural and children s hospitals to maintain access to care for Medicaid recipients. In addition, Florida Medicaid applied multipliers that increase the DRG base rates for long term acute care and specialty rehabilitation hospitals to improve upon inequities in payment that had evolved over time. Arizona Medicaid chose to use a separate base rate for freestanding specialty hospitals, such as rehabilitation and surgical specialty hospitals, to maintain consistency with Arizona s pre-drg payment method. Page 25 December 29, 2017

30 Arkansas DRG Conversion Plan 6.2 Standardized Base Rates with Wage Area Adjustments Standardized Base Rates with Wage Area Adjustments - Discussion Alternatively, the approach used by many states and by the Medicare program is to establish a single system-wide standardized amount, and adjust the standardized amount by each hospital s geographic wage area index or factor. The wage areas and associated wage indices can be state-defined values or can be linked to the Medicare values. Adjustment by wage area allows for higher payment in geographic regions that have historically reported higher wage rates for hospital employees. Wage area indices act as multipliers to standardized base rate(s) and can be applied either to the entire base rate or to a portion of the base rate. For example, Medicare applies the wage area index only to a percentage of the standardized base rate where the percentage is a predetermined estimate of the percentage of hospital costs attributed to labor. In particular, Medicare applies the wage index to 62.0 percent of the common base rate for hospitals with a wage index less than 1 and applies the wage index to 68.8 percent of the common base rate for hospitals with wage index greater than or equal to 1. For example, the base rate for a hospital with a wage index greater than 1 is: Base rate = ([Common base rate] * [hospital wage index] * 0.688) + ([Common base rate] * 0.312) Medicare wage indices in Federal Fiscal Year (FFY) 2018 for providers located in Arkansas Medicaid range from to and the average is The difference from the lowest wage index to the highest is which is just under 20 percent of the average (see Appendix C). An alternative to adopting Medicare s wage indices would be to develop Arkansas-specific wage indices. However, determination of wage areas can be very complicated and would likely require DHS to take on a significant amount of administrative effort. 6.3 Funding for Provider Base Rates Funding for Provider Base Rates - Discussion Of the $857M in 2016 Arkansas Medicaid inpatient expenditures (per our models), $582M was in the form of claim-based payments, and $275M was in the form of supplemental payments. To develop a budget neutral DRG-based payment system where aggregate modeled payments under DRGs are equal to payments under the current system the starting point for the DRG funding pool is the claim payments under the current per diem system. For modeling purposes (as shown in Chapter 9 and in Appendix A), we have used the claim allowed amounts in the CY 2016 inpatient claims data to determine the claim-based funding pool. For Arkansas Medicaid, however, there are significant funding streams outside of the claimbased payment pool in the form of supplemental payments. Much of the supplemental payment funding is based on cost-settlement payments and Access payments designed to reimburse provider classes up to the Medicaid Upper Limit ( UPL ). Supplemental payments under the current system that are include in our DRG payment simulation modeling are as follows: Advance Tentative IP s Page 26 December 29, 2017

31 Arkansas DRG Conversion Plan Tentative Adult Expansion s Children's ACH Pediatric UPL s Hospital IP Access UPL s Public IP UPL s Residential Treatment Unit s The State may consider incorporating supplemental payment funding into the DRG payment system in the form of higher base rates (and by extension, higher claim-based payments). As discussed further in Chapter 9.3 and in shown in Appendix A, this shift in funding from lump sum payments based on cost-shortfalls to claim payments based on hospital case mix may result in significant fiscal impacts as supplemental payments are redistributed across providers under DRGs. 6.4 Per Diem Base Rates Per Diem Base Rate - Discussion As mentioned previously, some provider types and some types of services may be carved out of the DRG payment method because they are more appropriately paid via another method. If such a decision is made, the carved-out services will presumably be paid per diem as that is the current DHS inpatient payment method, and per diem rates will need to be determined. The current method used to create per diem rates may be acceptable, in which case no changes need to be made. However, the current method may be unnecessarily cumbersome when applied to only a relatively small subset of inpatient stays, and, if so, DHS may want to consider adjusting the per diem rate setting process. Options for setting per diem base rates include setting rates based on average hospital cost per day and using a graduated scale based on length of stay as Medicare uses for paying psychiatric services. In addition, the availability of DRG grouping allows the option of calculating case mix adjusted per diems, similar to the way Medicare pays for some services. Furthermore, for a limited number of specialty services, a percent of charges (cost based) method could be considered in place of a per diem payment method. Page 27 December 29, 2017

32 Arkansas DRG Conversion Plan 7 Pricing Logic 7.1 Pricing Flow Figure 7.1 below shows the basic flow of DRG pricing logic. This flow may need to be customized slightly depending on the final DRG payment method design for Arkansas Medicaid. Figure 7.1: DRG pricing flow Determine DRG code Calculate outlier payment amount Retrieve DRG relative weight, average length of stay and policy adjustors Adjust payment for non-covered days Retrieve hospital base price Adjust payment if Medicaid allowed amount is greater than submitted charges Calculate base payment = [hosp base price] * [DRG rel wt] * [policy adjustors] Calculate final Medicaid allowed amt = [DRG base pymnt] + [outlier amt] Adjust DRG base payment for acute-to-acute transfers Calculate reimb. amount = [allowed amount] [Oth ins] [Spend down] [Pat Res] DRG codes, DRG relative weights, and hospital base rates were discussed previously in Chapters 5 and 6. The following sections of this chapter discuss the rest of the factors involved in calculating a DRG-based price. 7.2 Policy Adjustors Policy Adjustors - Discussion Policy adjustors are multipliers applied to specific claims for the purpose of increasing or decreasing payment. Generally, policy adjustors are applied for specific types of care, either for all recipients receiving that care or for subsets of recipients. Four types of policy adjustors are commonly used: Page 28 December 29, 2017

33 Arkansas DRG Conversion Plan Service adjustors Age/service adjustors Provider/service adjustors Provider adjustors Policy adjustors are an optional feature that can be used to help protect access to care for specific services. Often these are used for services where Medicaid funding can have a significant impact on beneficiary access, such as obstetrics, newborn care, mental health and pediatrics. The adjustors are above and beyond DRG relative weights and represent an explicit decision to direct funds to a particular group of patients who are otherwise clinically similar. Also, assuming a goal of budget neutrality, use of policy adjustors cause hospital base rates to be reduced having the effect of shifting some money from one area to another. We generally recommend including policy adjustor functionality in a DRG implementation because it creates an ability to meet current and future Medicaid program goals by adjusting payments without requiring significant software changes within the MMIS. However, policy adjustors do not necessarily need to be a major contributor to overall program reimbursements. They can be used sparingly to meet specific needs. The first type of policy adjustor, service adjustor, works particularly well if there is a desire to increase payment for specifically targeted services, such as obstetrical and neonatal care. The age/service adjustor is better suited if DHS desires to adjust payment for recipients within specific age categories, such as adjusting all pediatric services or adjusting for services provided to recipients under the age of one. Age/service adjustors provide a different payment for similar services when provided to one age category versus another. For example, a pediatric age/service adjustor of 1.25 on APR DRG (pneumonia severity 1) would increase payment by 25 percent if the patient was a child. In contrast, an adult whose claim mapped to APR DRG (pneumonia severity 1) would receive the DRG base payment without any adjustment. In truth, age/service adjustors can be applied to any age range, but are typically used by Medicaid programs to promote access for pediatric beneficiaries. Provider/service adjustors can be used to increase (or decrease) payment for specific services when offered by specific groups of providers. For example, a Medicaid agency might choose to increase payment for neonatal care when offered at a specialty children s hospital which might incur greater costs to support clinical expertise and equipment needed to treat very sick children. In such a scenario, a provider/service adjustor could be used to increase payment for neonatal care when provided at children s hospitals without increasing payment for other types of care (such as normal deliveries) at the same hospitals. Finally, provider adjustors can be used to increase (or decrease) payments for all services performed by specific individual providers or categories of providers. Provider adjustors differ from provider/service adjustors in that they apply for all stays at a particular hospital, not just stays for certain types of services. Provider adjustors serve the same purpose as applying different DRG base rates to different categories of hospitals, as they apply to every discharge for Medicaid recipients at the applicable hospitals. Even so, use of provider policy adjustors is sometimes politically more palatable than assigning different DRG base rates to different categories of hospitals. Page 29 December 29, 2017

34 Arkansas DRG Conversion Plan Within DRG pricing calculations, the adjustors can affect the DRG base payment using the following formula: [DRG Base ] = [Hospital Base Rate] * [DRG Relative Weight] * [Service Adjustor] * [Age/Service Adjustor] * [Provider/Service Adjustor] * [Provider Adjustor] Or only the highest adjustor can be used, in which case the formula is, [DRG Base ] = [Hospital Base Rate] * [DRG Relative Weight] * Maximum of ([Service Adjustor], [Age/Service Adjustor], [Provider/Service Adjustor], [Provider Adjustor]) For any particular service, one, two, three, or all four of the adjustors can be, and very commonly are, set to 1.0, thus creating no adjustment. The types or categories of service for which policy adjustors are applied are identified by DRG codes. Each DRG code is assigned a DRG relative weight and three adjustor values: service, age, and provider. In theory, a Medicaid program could simply make adjustments to DRG relative weights outside the MMIS and avoid putting separate adjustor fields into the MMIS. However, this would upset the integrity of the DRG relative weights and is something we strongly discourage. DRG relative weights are intended to indicate relative hospital resource utilization and patient acuity, and can be used to measure hospital case mix. Those measurements would not be valid if the DRG relative weights were manipulated. 7.3 Transfer Adjustments Transfer Adjustments - Discussion DRG payments are designed to be a single payment for a complete stay in a hospital. Given this design, full DRG payments can be unnecessarily high if a patient is transferred from one acute care facility to another resulting in an unusually short length of stay at the transferring from hospital. To handle this situation, most Medicaid DRG implementations have followed the Medicare model in which a payment amount is calculated using a per diem method and then compared to the DRG base payment. The per diem payment is referred to as a transferadjusted base payment amount and, if less than the DRG base payment, is used in place of the DRG base payment. The formula used to calculate the transfer-adjusted base payment is: Transfer Adjusted Base Pymt = {[DRG Base ] / [DRG Average Length Of Stay]} * {[Length Of Stay] + 1} Adding one to the length of stay takes into account the disproportionate amount of costs required in the first day of admission to complete the admission process and perform an initial diagnostic evaluation. Under this particular formula, the transfer adjusted base payment comes out less than the DRG base payment if the length of stay is less than the DRG s average length of stay minus 1. Otherwise, the transferring from hospital receives full DRG payment. For average length of stay data, DHS can use arithmetic or geometric averages from untrimmed or trimmed national data as published by 3M along with the national weights. In addition, statewide averages can be used, or national averages calculated using data from the Nationwide Inpatient Sample. Page 30 December 29, 2017

35 Arkansas DRG Conversion Plan Transfer payment adjustments only apply to the transferring hospitals. Receiving hospitals are paid the full DRG amount. The transfer payment adjustment process is used when a patient is transferred from one acute care setting to another. Transfers are identified in claims data through the discharge status and DHS s DRG payment policy will need to specify which discharge status codes apply to the transfer payment adjustment process. Possible status codes to include are: 02 discharged/transferred to a short-term general hospital for inpatient care 05 discharged/transferred to a designated cancer center or children s hospital 07 left against medical advice (Medicare uses this value if the patient is admitted to another acute care hospital on the same day) 43 discharged/transferred to a federal facility 62 discharged/transferred to an inpatient rehabilitation facility or distinct part unit 63 discharged/transferred to a long-term care hospital 65 discharged/transferred to a psychiatric hospital or distinct part unit 66 discharged/transferred to a critical access hospital 82 Discharged/transferred to a short term general hospital for inpatient care with a planned acute care hospital inpatient readmission 85 Discharged/transferred to a designated cancer center or children s hospital with a planned acute care hospital inpatient readmission 93 Discharged/transferred to a psychiatric hospital/distinct part unit of a hospital with a planned acute care hospital inpatient readmission 94 Discharged/transferred to a critical access hospital (CAH) with a planned acute care hospital inpatient readmission DHS may also consider a post-acute care transfer policy similar to that used by Medicare. This policy reduces payment to hospitals for a specified list of DRGs (280 MS-DRGs in FFY 2018) when the patient is transferred to a particular type of hospital. The need for this policy arose from the disparate payment incentives facing acute care providers (paid per stay) and post-acute care providers (paid per day). For patients requiring both acute and post-acute care (as identified by the list of 280 MS-DRGs, for example, hip replacement), Medicare reduces payment to the hospital if a stay is particularly short and the patient is discharged to a postacute setting. Patient discharge status codes that Medicare includes in its post-acute care transfer policy are: 03 discharged/transferred to a skilled nursing facility 05 discharged/transferred to a cancer or children s hospital 06 discharged/transferred to a care of a home health agency (exceptions apply) 62 discharged/transferred to a rehabilitation facility or distinct part unit 63 discharged/transferred to a long-term care hospital 65 discharged/transferred to a psychiatric hospital or distinct part unit 83 discharged/transferred to a skilled nursing facility with a planned acute care hospital inpatient readmission 85 discharged/transferred to a cancer center or children s hospital with a planned acute care hospital inpatient readmission 86 discharged/transferred to a care of a home health agency with a planned acute care hospital inpatient readmission (exceptions apply) 90 discharged/transferred to a rehabilitation facility or distinct part unit with a planned acute care hospital inpatient readmission Page 31 December 29, 2017

36 Arkansas DRG Conversion Plan 91 discharged/transferred to a long-term care hospital with a planned acute care hospital inpatient readmission 93 discharged/transferred to a psychiatric hospital or distinct part unit with a planned acute care hospital inpatient readmission 8 Medicare has a large enough percentage of their population fitting this scenario to justify incurring the extra administrative complexity of this post-acute transfer policy. With the possible exception of transfers to psychiatric hospitals, Medicaid programs have a significantly lower percentage of their populations fitting this scenario, so the added complexity of this policy may be unwarranted. 7.4 of Non-General Revenue Funds Distributed on a Claim-by- Claim Basis of Non-General Revenue Funds Distributed on a Claim-by-Claim Basis - Discussion As mentioned previously, funds from inter-governmental transfers (IGTs) and hospital assessments make up a significant portion of the total reimbursements paid out through the Arkansas Medicaid program. Shifting some of these reimbursements away from periodic lump sum distributions and to the claim level payments will satisfy current CMS policy and will help facilitate a migration of recipients into the PASSE. Shifting payments to hospital claim payments can be done by making the funds available to increase the DRG base rate, or by paying per-claim-add-on payments that are separate from the DRG base rate. Adding the funds to the DRG base rate has the advantages of making outlier calculations more straightforward on individual claims and allowing for payment on individual claims to be increased or decreased based on the DRG relative weight. Whereas making the IGT/assessment-funded payments a per-claim-add-on, separate from the DRG base rate, has the advantage of making the accounting of these funds more straight forward. For example, Florida Medicaid initially implemented its DRG payment method with two different per-claim add-on payments each with separate annual distribution amounts for different hospitals. Alabama Medicaid planned a flat percentage (about 12 percent) of the DRG payment to be distributed as a separate add-on payment. This option allowed the per-claim-add-on amount to be affected by patient acuity while still maintaining separate accounting of these funds Other Per Claim Add-On s Other Per Claim Add-On s - Discussion In addition to distributing non-general Revenue Funds (GRF), some DRG installations include per-claim add-on payments, for other purposes. For example, Medicare offers per-claim add-on payments for direct graduate medical education costs (Medicare also provides payment adjustments for indirect medical education costs, capital, and disproportionate share hospitals, 8 MLN Matters Number: SE0801, reissued on November 17, 2015; downloaded from MLN/MLNMattersArticles/downloads/SE1411.pdf on October 12, The Alabama Medicaid implementation of DRG payment is currently on hold pending decisions regarding implementation of Medicaid managed care. Page 32 December 29, 2017

37 Arkansas DRG Conversion Plan but these adjustments are made to the common base rates 10 ). Montana Medicaid provides separate add-on payments for medical education, capital, and disproportionate share payments. Similarly, Washington DC Medicaid provides per-claim add-on payments for medical education and capital. Other supplemental payments can also be distributed through add-on payments if distribution of the funds makes sense to be made on a per claim basis. 7.6 Outlier s Outlier s - Discussion In general, DRG payment is designed to pay based on average hospital resource usage, as is a per diem payment. DRG payment methods generally align payment with hospital resource usage slightly better than per diem methods. However, being based on averages, both methods will have cases in which payment is relatively low and cases where payment is relatively high. In general, adjustments to standard DRG payment should only be made when the difference between that payment and hospital cost is extreme. For cases of extreme under payment, DRG payment methods typically include outlier provisions that adjust payment upward for stays that are unpredictably expensive. DRG groupers are limited to using only the information on medical insurance claims including principal diagnosis, procedures, age, complications and comorbidities (identified through secondary diagnosis codes), and discharge status. Given the tremendously wide range of cases seen in an inpatient setting, DRG grouping, although continually improving, does not always accurately predict hospital resource use. In those cases, where the prediction differs significantly from reality, outlier payments are used to generate a more reasonable reimbursement. Medicare and many Medicaid agencies utilize a cost-based stop-loss model that applies outlier payments if the estimated loss to a hospital exceeds a dollar amount threshold. When the threshold is exceeded, remaining costs are reimbursed at some percentage. This percentage is referred to as a marginal cost factor because it is intended to cover only the marginal costs of the additional care. These costs include only variable costs such as staffing and supplies, not fixed costs such as buildings and equipment. Medicare s marginal cost factor is 80 percent (90 percent for burns) and states values range from 50 percent to 100 percent, and sometimes vary by the type of service. A variety of strategies are used to set the estimated loss threshold. Medicare uses a single threshold. California Medicaid has selected two thresholds, with one marginal cost percentage (60 percent) used for losses between threshold 1 and threshold 2 and a second marginal cost percentage (80 percent) applied for losses above threshold 2. Other states base the outlier threshold on the DRG relative weight, for example, Ohio and Washington, DC, while other states, for example Pennsylvania, set the outlier threshold to some percentage of the DRG base payment, such as 150 percent. Yet another example would be to set a small number of thresholds for different types of services so that less expensive services, such as mental health care, can be given a lower threshold than more expensive types of services. Under the cost-based stop-loss outlier payment model, a method has to be selected for determining cost-to-charge ratios (CCRs) for purposes of estimating hospital cost. A single state-wide CCR can be used, separate CCRs for each hospital can be determined one per hospital, or separate CCRs can be determined for each standard cost center for each hospital. 10 Medicare Learning Network (MLN), Acute Care Hospital Inpatient Prospective System System Fact Sheet, ICN , February Page 33 December 29, 2017

38 Arkansas DRG Conversion Plan The lower level of granularity in CCRs offers greater accuracy in estimating costs, but has the trade-off of requiring additional effort to periodically recalculate the values. Less commonly, outlier cases are identified by length of stay being above a threshold number of days. For days above the threshold a per diem amount can be paid to help alleviate hospital losses. Rhode Island and Mississippi, for example, use a length of stay outlier threshold for mental health stays. Including a day outlier adds complexity to the overall payment method, but provides a more equitable outlier payments for services like mental health care, which are generally less expensive, and are less likely to hit a single cost outlier threshold. Setting outlier threshold(s) and marginal cost percentage(s) are a policy decision. Generally, the values are set so that outlier payments are within a pre-determined range of total payments. For example, Medicare generally aims for an outlier payment percentage between 5 and 6 percent. Medicaid programs tend to have a slightly higher percentage of high-cost cases and generally aim for an outlier payment percentage between 5 and 10 percent. The percentage of payments made through outliers can be adjusted by increasing or decreasing the outlier threshold and/or increasing or decreasing the marginal cost percentage. As previously described in Chapter 3, a common formula used to calculate the outlier payment on a claim is: [Outlier Pymt Adjstmnt] = {[Hospital Cost] [DRG ] [Outlier Threshold]} * [Marginal Cost %] and outlier payments are only made if {[Hospital Cost] [DRG ]} is greater than the outlier threshold. simulations can be made in which the outlier threshold and the marginal cost percentage are adjusted until the desired outlier payment percentage is reached. Provider base rates and policy adjustors can also be manipulated resulting in an increase or decrease of total outlier payments. From a policy perspective outlier payments are important to ensure access to care for very high cost cases. Providers need to know they will be compensated if they treat very sick individuals. However, paying too much out in the form of outliers removes provider incentives to contain costs as outlier payments are cost based increasing when costs increase. In addition, in a budget neutral system, an increase in reimbursements paid out as outliers generates a reduction in provider base rates. These trade-offs are typically balanced in Medicaid programs by setting a target outlier payment in the range of 5 to 10 percent, and outlier threshold and marginal cost percentage are set to hit that target. A completely different strategy for dealing with outlier cases is to shift them out of the DRG payment method and pay them with some other method, such as percentage of cost or per diem. These methods may be more amenable to hospitals; however, they remove some of the incentives to control costs provided by a DRG payment method. They also complicate the overall Medicaid inpatient payment method because individual providers are reimbursed using more than one process. 7.7 Provider Gain Adjustments Provider Gain Adjustments - Discussion Most outlier cases are stays where the costs to the hospital far outweigh the payment, but the opposite also occurs where payment far exceeds hospital cost (this occurs most often with patients who expire). Medicaid agencies may implement a policy that adjusts payment Page 34 December 29, 2017

39 Arkansas DRG Conversion Plan downward in these scenarios. Such a policy avoids potential negative publicity for a rare occurrence in which DRG payment is very high in relation to the hospital s charges and/or cost of care. In addition, this type of payment policy has the benefit of shifting money, albeit a relatively small amount of money, from highly profitable stays into other stays. Downward payment adjustment may be implemented through a charge cap in which Medicaid allowed amount is capped at provider charges. A more sophisticated downward payment adjustment can be implemented using a provider gain cost outlier policy which works very much like the more common provider loss cost outlier policy. In a provider gain cost outlier policy, a provider gain threshold is set and a marginal cost factor is used to hold back some of the provider gain above the fixed gain threshold. 7.8 Short Stay Adjustments Short Stay Adjustments - Discussion The current per diem payment method provides a payment for each day a patient is in the hospital. Thus, hospitals can maximize payment by maximizing the number of days of a patient s stay. In contrast, a DRG payment method provides a payment for each hospital discharge (also can be thought of as payment per admission). Under a DRG payment method, hospitals can maximize payment in relation to cost by either limiting the length of stay, or where possible, maximizing the number of admissions/discharges. Some Medicaid agencies implement policies to reduce payment for very short stays to reduce this incentive and help ensure appropriate care is provided to each patient. One potential payment policy that reduces this incentive is the provider gain outlier policy mentioned in the previous section. Another possibility implemented by some Medicaid agencies is a short-stay payment reduction policy. For example, Blue Cross and Blue Shield of Tennessee has a short stay policy that pays applicable claims via a per diem. To qualify as a short stay, the assigned APR DRG must have an average length of stay greater than five days and the actual length of stay for the hospital admission must be less than or equal to 20 percent of the DRG average length of stay (rounded down) Non-Covered Days Adjustments Non-Covered Days Adjustments - Discussion As mentioned in an earlier section related to transfer claims, a DRG payment is designed to be a single payment for a complete hospital stay. This kind of payment will be inappropriate if the recipient did not have Medicaid fee-for-service eligibility and benefit coverage for the entire stay. If some of the days of a stay are not covered then a reduction should be applied to the full DRG payment. Having eligibility for only part of a hospital stay is relatively rare in a Medicaid program, but can happen at times either because a recipient lost or gained Medicaid eligibility during the hospital stay or shifted from fee-for-service to managed care during the stay. In addition, some recipients have benefit coverage only for emergency services. If these recipients are deemed to be in an emergency medical condition for part, but not all of an inpatient stay, then the Medicaid payment should cover only part of the hospital stay. Thus, this scenario is very similar to a partial eligibility scenario. Recipients who are eligible only for emergency services include undocumented non-citizens. 11 Hospital Contract Modernization: Business Rules presentation dated May 11, 2011 and downloaded from on October 12, Page 35 December 29, 2017

40 Arkansas DRG Conversion Plan One option for reducing payment in this scenario is to perform calculations very much the same as those used with transfer claims. A per diem type of payment, referred to as the non-coveredday adjusted base payment, can be calculated and compared against the full DRG base payment. If the eligibility-adjusted base payment is less, it can be used in place of the full DRG base payment. Another option would be to prorate the full DRG payment based on the number of covered days. For example, if a recipient is Medicaid fee-for-service eligible for 6 days out of a 10-day hospital stay, payment could be reduced to 60 percent of the full DRG payment. Similarly, if the recipient was covered only for emergency services and the recipient was deemed to be in an emergency medical condition for only 6 days of a 10 day stay, then payment could be reduced to 60 percent of the full DRG payment. Page 36 December 29, 2017

41 Arkansas DRG Conversion Plan 8 Other Considerations 8.1 Admit versus Discharge Date Admit versus Discharge Date - Discussion Initial implementation of DRG pricing and subsequent updates to DRG version and rates need to be implemented on specific dates that can be communicated to all interested stakeholders, including Medicaid policy, Medicaid fiscal agent, hospitals, hospital industry software developers, and PASSE plans. These implementation and version update dates are most commonly defined as date-of-service cutovers. That is all claims with a date of service prior to the implementation date are processed one way and all claims with a date of service on or after the implementation date are processed a different way. When considering hospital claims which usually cover multiple dates of service, the DHS will need to determine whether to apply the date of admission or the date of discharge to the implementation/upgrade date. 8.2 Transitional Period Transitional Period - Discussion Making a change in payment method from per diem to DRGs has potential to result in significant redistribution of funds, especially if supplemental payment funding is included in the DRG payment system. Even if implemented with budget neutrality in aggregate for the Medicaid program, we expect some hospitals will receive higher payments under the new DRG method (when compared to legacy per diem payments) and some hospitals will receive lower payments. Such changes in payments are common in these types of transitions. Some payers have established transitional policies to mitigate the impacts of such payment changes in the years immediately following implementation of a new DRG model. For example, when Medicare implemented DRGs for the first time, it provided a phase-in period of four years for the operating component of the new payment rates, and a 10-year period for the transition of the capital-related component of the rate. Some Medicaid programs, New York, Wisconsin, Arizona and California for example, have also used transition periods. On the other hand, other Medicaid programs, including those in Pennsylvania, Washington and Kentucky, have not provided for phase-ins or transitional periods. Similarly, when Medicare transitioned from the legacy CMS-DRG model to its new severity-based MS-DRG model, it did not use a transition period. There are some advantages to utilizing transitional strategies. Phase-in or transitional periods provide time for providers to internally respond to anticipated changes in Medicaid funding. A transitional period allows time for providers to take the steps necessary to improve documentation and coding practices, and potentially implement improvements to operating performance relative to efficient delivery of services. In addition, a transition period gives hospitals time to make modifications to the complement of service lines offered in future periods to the extent that Medicaid payments affect such decisions. On the other hand, there are disadvantages to utilizing transitional strategies. From a payer perspective, transitional periods tend to increase the program administrative complexity of both policy implementation and system implementation. It also requires payers to either maintain two payment systems simultaneously (which would be required to blend payments between a per diem and DRG model), or alternatively, to determine hospital-specific base rates that would Page 37 December 29, 2017

42 Arkansas DRG Conversion Plan effectively build in the transition to such rates. From the providers perspective, hospitals that stand to see increased payments under the new payment model will not realize the full benefit of the change in payment model until after the transition period has run its course. Mechanically there are two main approaches to applying transitional adjustments to claim payments, as follows: Claim payment adjustments to limit gains or losses to a floor or ceiling (only to hospitals beyond impact threshold, outside of gain/loss corridor). For example, the State could limit modeled payment increases to 5 percent and modeled payment decreases to -5 percent, with additional adjustments to make aggregate system payment payments budget neutral. Under this approach, hospitals within the corridor (in between the ceiling and floor) would not have an adjustment. Claim payment adjustments based on a percent of modeled payment change for all providers. For example, the State could adjust each provider s payment such that they have 50 percent of the modeled gain or loss. If a transition period is established by Arkansas Medicaid, decisions will need to be made regarding whether the transition will be budget neutral or will include new money that will be available for some finite amount of time. Decisions also need to be made regarding the length of the transition, generally between one and three years, and the method of implementing the transition. For example, during Florida Medicaid s implementation of DRG pricing, the Florida Legislature allocated $65 million in new money for one year to help offset hospital losses from DRGs. This money was distributed to hospitals in three tiers 1) rural hospitals were given enough funds to cover any losses from the shift to DRG rates; 2) hospitals that experienced more than 10 percent or higher reductions in Medicaid inpatient reimbursements were given enough transitional funds to cover 24 percent of their payment reductions; 3) hospitals experiencing less than 10 percent reduction in Medicaid inpatient reimbursements were given enough transitional funds to cover 4 percent of their payment reductions. In addition for non-rural hospitals, the hospital needed a Medicaid inpatient payment reduction of at least $300,000 from the shift to DRG pricing to qualify for any of the transitional funds. Thus, the one year of available new money helped mitigate some, but certainly not all, of the payment reductions experienced by individual hospitals. 12 As another example, California Medicaid (Medi-Cal) implemented a budget neutral transition period that lasted for three years. Medi-Cal s transition policy limited both increases and decreases in payments to individual hospitals for the first three years of DRG pricing. In year one, hospitals received a base price that was within 5 percent of the base price that would maintain their funding level compared to the prior methodology assuming the same patient acuity and volume of stays. This increase or decrease limit was set to 10 percent in year two and 15 percent in year three. In year four, all hospitals that have not already done so will transition to their appropriate base price. Hospitals with very low Medi-Cal volume did not qualify for the transition period Diagnosis Related Group Transitional s report submitted by the Florida Agency for Health Care Administration to the Florida Legislature on June 30, Medi-Cal DRG Executive Summary; downloaded from on October 12, Page 38 December 29, 2017

43 Arkansas DRG Conversion Plan 8.3 Documentation and Coding Adjustment Documentation and Coding Adjustment - Discussion Under a DRG payment method, overall case mix (i.e. average DRG relative weight) has a significant impact on overall Medicaid payments. This can be seen when looking at the DRG base payment formula: [DRG Base ] = [Hospital Base Rate] * [DRG Relative Weight] * [Policy Adjustor(s)] While payments under a DRG payment method are also affected by policy adjustors, outlier payments, and transfer and non-covered days adjustments, these additional factors all have a relatively small impact on overall spending when compared with the impact resulting from changes in case mix. The significance of potential changes in case mix relative to overall Medicaid spending for inpatient hospital services punctuate the need to accurately estimate these values and to monitor them through the first few years of a DRG payment implementation. If case mix is significantly understated during the design process, resulting Medicaid spending will likely be well above estimates. When considering the increases in case mix that will occur after implementation of the APR DRG model, there are generally two components. One component is attributable to actual changes in the patients health status, where hospitals are required to expend more resources because patients they are treating are actually sicker this component is commonly referred to as a real increase in acuity. And there is an expectation that case mix will increase slightly from year to year, all other things remaining equal. As an example, before Medicare s implementation of MS-DRGs in 2008, annual case mix increases ranged from -0.8 percent to 1.0 percent, and on average reflected 0.1 percent year-to-year change. 14 This slight increase may be the result of a number of factors, including the trend of providing more and more services efficiently and effectively in outpatient settings, leaving only sicker patients in the inpatient hospital setting. Increases can also be attributable to advances in medical technology that allow hospitals to be more effective in caring for the sickest of patients. A DRG system is generally designed to self-adjust for this type of case mix increase as patients get sicker, they are classified into DRGs with higher relative weights, and as a result, payments for services increase. Payers must set rates appropriately, considering these small increases in case mix over time. The other component of the increase in case mix can generally be attributable to documentation and coding improvements (DCI). During DHS s transitions from the legacy per diem payment model to a new APR DRG per discharge payment model, there is an expectation that DCI will result in the rate of increase in reported case mix being higher than it would have been if DHS had decided to maintain the current legacy per diem model. This increase in case mix can be attributed to improvements in medical record documentation and improvements in claim coding as Medicaid claim payment becomes dependent on the diagnosis and procedure codes on claims. These documentation and coding improvements are an appropriate and necessary response by providers to DHS s implementation of the APR DRG payment model. 15 However, 14 Medicare Advisory Commission (MedPAC), Report to the Congress: Medicare Policy (March 2011), p Expected increases in case mix resulting from DCI should not be confused with case mix increases attributable to the term DRG creep, which may be the result of inappropriate billing practices. Such inappropriate billing practices are intended to generate higher payment by billing for services in a way that provides for a higher payment rate (DRG Page 39 December 29, 2017

44 Arkansas DRG Conversion Plan the underlying cause of this component increase is very different. It is due to better reporting, not to actual changes in types of patients treated. Also, the increase in case mix from DCI is expected to be much more significant than real case mix change, when comparing the first year of DRG implementation to the simulation dataset containing claims that were billed and paid under a per diem method. There is a risk that the case mix values reflected in the simulation data significantly understate the actual acuity of the patients served. This understatement is primarily the result of coding and documentation practices that were intended to support payment under the legacy per diem model, and not intended to support payment under an APR DRG model. The coding practices that are necessary to generate accurate payments under the two models are significantly different, with the standard under an APR DRG model being much higher. The potential impact of this DCI on case mix is illustrated in Figure 8.3 below, using hypothetical values. As shown in the following illustration, it is our expectation that there will be an immediate bump in aggregate paid case mix in the first year following implementation of the new system, with smaller increases in following years, and with paid case mix increases trending back to pre-implementation levels once providers successfully improve their coding and documentation processes. Figure 8.3 Illustration of Potential Impacts to Paid Case Mix from DCI Clearly, a potential financial risk to both the State and to the providers exists as a result of this situation. Understanding that the DRG payment rates that will be implemented will be based on payment simulation models that reflect a potentially understated case mix, the State will be at risk of overspending its budget in the event that actual case mix exceeds expected levels after the system s effective date. On the other hand, the providers are at risk if the opposite is true that the simulation models used to set DRG payment rates overstate actual case mix, although with higher relative weight) than what accurately reflects the condition and treatment of the patient, oftentimes accomplished through miscoding or inappropriate re-sequencing. Page 40 December 29, 2017

45 Arkansas DRG Conversion Plan this scenario is less likely. In addition, providers are at risk if the State over compensates for anticipated changes in case mix. The challenge related to this issue is to implement a strategy that effectively mitigates the potential risk of overpayment (or underpayment) to both the State and to the providers. Other government payers have experienced significant increases in paid case mix following system changes, as did the Medicare program when it transitioned to the MS-DRG payment model. In its March 2011 Report to the Congress, the Medicare Advisory Commission (MedPAC 16 ) reported that the implementation of Medicare s MS-DRG model gave hospitals a financial incentive to improve medical record documentation and diagnosis coding to more fully account for each patient s severity of illness. While documentation and coding improvements (DCI) appropriately improve measurement of patient severity, they also can increase reported case mix under MS-DRGs even if patients levels of illness and resource need are not different from prior years. The result was strong growth in payments per case in 2008 and Analysis by CMS found (and MedPAC s analysis concurred) that payments increased by a total of 5.8 percent over the two years due to coding improvements. 17 MedPAC s report characterized this increase as extraordinary since it followed a decade in which the case-mix index declined in 5 of the 10 years and never grew by more than 1 percent in any year. 18 At the state payer level, Florida Medicaid s experience provides another recent example of the potential impact on paid case mix resulting from a transition to APR DRGs. The Florida Medicaid program implemented an APR DRG payment system on July 1, Similar to Arkansas, Florida converted from a per diem payment method to a payment method based on APR DRG categorizations. Having information regarding experience from other state Medicaid DRG implementations, in its first year of DRG pricing, Florida Medicaid decided to plan for a case mix change of four percent from documentation and coding improvement and one percent from real case mix change, resulting in a total case mix adjustment of five percent. The real case mix adjustment was based on an assumed real case mix increase equal to one-third of one percent per year. The historical claim data used for rate setting included claims with dates of admission three years prior to year 1 of DRG pricing, resulting in a total of one percent real case change. To reduce risk for both the Medicaid agency and the hospital community, Florida Medicaid also included a provision to perform a mid-year adjustment if the assumptions regarding case mix change proved inaccurate. This provision provided authority to increase or decrease DRG base rates for the last quarter of the state fiscal year based on actual case mix experience measured on claims paid under the DRG payment method. Florida Medicaid continued this basic strategy of assuming significant case mix change resulting from DCI and providing authority for mid-year payment adjustments for the first three years of their DRG implementation. Starting in year four, the historical claim data used for rate setting included claims billed and paid under DRG pricing. From that point forward, Florida Medicaid s DRG rate setting process no longer assumes any case mix change from DCI. 16 MedPAC is an independent Congressional agency established by the Balanced Budget Act of 1997 (P.L ) to advise the U.S. Congress on issues affecting the Medicare program. The Commission's statutory mandate is quite broad: In addition to advising the Congress on payments to private health plans participating in Medicare and providers in Medicare's traditional fee-forservice program, MedPAC is also tasked with analyzing access to care, quality of care, and other issues affecting Medicare. 17 MedPAC Report to Congress (March 2011), pages 39-40, emphasis added 18 MedPAC Report to Congress (March 2011), page 49 Page 41 December 29, 2017

46 Arkansas DRG Conversion Plan The actual case mix in the first three years of DRG pricing for Florida Medicaid was higher than the case mix on the historical model claims used for rate setting as reflected in Table 8.3. Table Case Mix Change Between Modeling Period and Claims Paid Under DRGs Claim SFY Historical Model Data Used for Rate Setting Increase in Case Mix Since Modeling Period SFY 2014 SFY % SFY 2015 SFY SFY 2016 SFY % Given the experience in year one (SFY 2014), Florida Medicaid increased their DCI estimate to a total of seven percent in years two and three (SFY 2015 and SFY 2016). Even with this change, actual case mix was higher than predicted and mid-year rate adjustments were performed in both years two and three. Mid-year rate adjustment was not performed in year one because actual case mix increase was not measured to be higher than five percent until the end of the fiscal year (at which time it was too late to make adjustments). Florida Medicaid used a date of admission cut-over to DRG pricing and the lag between date of admission and the time case mix could be measured on paid claims proved to be too long to allow for a mid-year base rate adjustment in year one. Given the risk of payment increases driven by DCI, there are several options available to DHS as strategies to mitigate potential over or underpayment of services. These options are: Option 1: Prospectively reduce either base rates or relative weights to reduce future payments to offset anticipated increases in payments resulting from DCI. This is generally the approach that was taken by CMS when it implemented the MS-DRG payment system for Medicare services. The key challenge with this option is accurately estimating in advance what the increases related to DCI will be in future periods. Option 2: Retroactively adjust either base rates or relative weights to offset actual increases in payments resulting from DCI. To implement this option, it would be necessary to first estimate what expected, or real case mix increases should be, based on historical trends. To the extent that actual case mix increases exceed the established case mix increase trend line, adjustments can be made. Adjustments could be made in the form of retroactive adjustments of historical claims (e.g., through mass adjustments to claims) or through reductions to future payments. Option 3: Establish a hybrid strategy that establishes a prospective adjustment with a corridor (for example, the expected case mix in future periods based on historical trends, plus or minus a fixed percentage). Using this corridor, monitor actual paid case mix on a regular basis, and if it remains within the established corridor, make no adjustment going forward. If it falls outside of the established corridor, make an adjustment, either prospectively or retrospectively, to bring payments to where they would have been had the actual paid case mix not exceeded the upper bound of the corridor (or the lower bound of the corridor in the instance of a measured case mix reduction). There are a number of variations that can be applied to each of these options. Page 42 December 29, 2017

47 Arkansas DRG Conversion Plan 8.4 Interim Claims and Late Charges Interim Claims and Late Charges - Discussion DRG payments are designed to be single payments for complete hospital stays. Thus, a final DRG payment reflecting all diagnoses and procedures cannot be determined until the patient is discharged. For most hospital stays, that is perfectly acceptable to both the provider and the Medicaid agency. However, for very long stays, waiting until discharge for payment from Medicaid can cause cash flow challenges for hospitals. This can be solved by allowing interim billing and payment. Unfortunately, generating final payment for a hospital stay after interim payments have been made can be a challenging task to implement in an MMIS and adds complexity to the overall DRG payment method. As a result, decisions related to interim claim payments are an important part of a DRG payment policy despite the fact that they affect a relatively small percentage of overall stays. Current DHS allows for interim billing, and based on our review of the DRG model data, providers utilize interim billing extensively. Under the most extreme example, a provider submits separate interim claims for each individual day within the patient stay. This frequency of interim billing is not administratively feasible under per discharge APR DRG payment and will need to be reduced significantly. If DHS decides to continue to allow interim payments, then a series of design decisions must be made. First, the threshold minimum number of days per interim claim must be decided most states have selected 30 days when interim claims are accepted, although some states have made the limit as high as 180 days. Next the method of payment for interim claims must be determined. A per diem payment is used by some states, but is a very complicated option to implement in an MMIS. Another option is to pay the full DRG amount on the first interim claim, and then require the hospital to submit an adjustment of the original claim when submitting any subsequent interim claims. For example, if the length of stay limit for interim claims is 30 days, the first claim interim claim will be submitted with a length of stay of 30 days. If the patient is still in the hospital after 60 days, the hospital would submit an adjustment to the original claim that contained dates of service from date of admission to day 60. Medicaid would then recalculate DRG payment. Assuming the DRG assignment was the same on both claims, the adjusted claim would only include additional payment if the claim reached outlier status. Late charges (claims with bill type 115) are also problematic in a DRG payment method. To accurately calculate DRG payment, including outlier payments, all charges for the hospital stay need to be submitted on a single claim. For this reason, late charges are typically not accepted by Medicaid agencies paying via DRGs. 8.5 Medicare Crossover Comparison Pricing Medicare Crossover Comparison Pricing - Discussion Many Medicaid programs have implemented Medicare crossover comparison pricing logic. This logic is applied specifically to Medicare crossover claims and compares the Medicare allowed amount to the Medicaid allowed amount. It then sets Medicaid reimbursement amount so that the total provider reimbursement, combining Medicare and Medicaid payments, reaches the lower of the two allowed amounts. If DHS uses this kind of pricing logic, then Medicare Page 43 December 29, 2017

48 Arkansas DRG Conversion Plan crossover claims will need to be processed through the new DRG pricing method so that a DRG-based Medicaid allowed amount can be determined. Page 44 December 29, 2017

49 Arkansas DRG Conversion Plan 9 DRG Simulation Model 9.1 DRG Model Analytical Dataset DRG Model Analytical Dataset - Discussion To estimate the fiscal impact of a new APR DRG payment system relative to the current inpatient per diem payment methodology and the current supplemental payment programs, we developed a DRG payment simulation model using CY 2016 Arkansas Medicaid inpatient FFS claims data. The DRG payment simulation model analytical dataset consists of inpatient FFS claims data, excluding crossover claims (for Medicaid-Medicare dual eligibles) provided by DHS and extracted from the MMIS. 19 We used the following steps to determine the net claim records utilized in the DRG payment simulation model analytical dataset: Consolidated claims billed on an interim basis: Claims data provided by DHS included multiple claims billed on an interim basis (the MMIS currently does not create a final consolidated claim record for all services provided during an inpatient stay). APR DRG pricing requires a single claim record for each admission; as such we consolidated interim billed claims into a single record for each admission for modeling purposes. We identified interim claims using a combination of claim attributes resulting in a claim key to aid in consolidation: recipient ID, provider ID and date of admission. When consolidating interim claims into a single record for each admission, we incorporated all unique data elements (e.g. diagnosis codes, procedure codes, date range, etc.) and summed all days, charges and payments to reflect the full inpatient stay visit period. Excluded claims with dates of discharge outside of calendar year 2016: Applied data filter to establish the most recently available and fully mature year coded under ICD-10. Excluded psychiatric residential treatment facility claims: One provider specialty type, Psychiatric Residential Treatment Facility, was excluded as directed by DHS; all other inpatient provider types were included. Excluded out of state providers not designated as a border provider: DHS designates all out-of-state hospitals located within 50 miles of the Arkansas state border that participate in the Arkansas Medicaid program as border providers. Border providers are essentially treated as in-state hospitals under the current payment methodology and were included in the model analytical dataset. For modeling purposes, all claims originating from out-of-state providers not designated as border were excluded. Excluded claims classified with an ungroupable APR DRG: Current billing requirements and MMIS infrastructure align with an inpatient hospital per diem approach, and in some limited instances, providers in the CY 2016 analytical claims data period did not report data sufficient to assign a valid APR DRG. The result of incomplete or conflicting information results in DRG designation Ungroupable. Reasons for an Ungroupable status include but are not limited to: invalid recipient gender, invalid recipient age, invalid primary diagnosis and insufficient data to group the claim. Claims initially classified as Ungroupable were evaluated to identify patterns and select claims were adjusted and 19 Inpatient hospital claims identified using the following category of services: 31 Inpatient Hospital, 33 Inpatient Psychiatric Under 21, 75 Pediatric Inpatient Hospital, 77 Rural Inpatient Hospitals, 87 Inpatient Arkansas Teaching Hospital, A1 Inpatient Hospital Transplant, CA Critical Access Hospital. Page 45 December 29, 2017

50 Arkansas DRG Conversion Plan re-grouped where appropriate. Claim adjustments used to assign valid DRGs include the following: If a newborn claim has a procedure code for a circumcision and the gender is captured as female (invalid recipient gender), the claim is adjusted to reflect male If a mother s date of birth is captured on a newborn claim (invalid recipient age), the claim is adjusted to use first date of service as the date of birth In the raw MMIS inpatient claims data we observed several thousand deliveryrelated claims where the reported primary diagnosis code was the delivery result (e.g. single live born) rather than the reason for admission (e.g. encounter for fullterm delivery). For these claims we re-sequenced the primary and secondary diagnosis codes and re-grouped in order to assign the appropriate delivery DRG. Excluded claims where allowed amount is equal to $0: When a claim is not denied but reflects a payment amount of $0, the claim was removed from the analytic dataset. Table 9.1 below summarizes the claim consolidation and exclusion process: Table 9.1 Net Claims Included in Analytical Dataset Description Claims Total Allowed Amount Raw Claims Received 339,907 $ 1,224,707,877 Less collapsed interim claims 154,896 0 Net Claims with Consolidated Interim Claims 185,011 $ 1,224,707,877 Less exclusion 1: Claims outside CY 16 analysis period 86,784 $ 586,029,839 Less exclusion 2: Psych residential treatment facility ,436,476 Less exclusion 3: Out of state non-border 915 7,949,401 Less exclusion 4: Ungroupable 85 1,476,784 Less exclusion 5: $0 allowed amount Net Claims Included in Analytical Dataset 96,114 $ 581,815,376 Page 46 December 29, 2017

51 Arkansas DRG Conversion Plan 9.2 Modeling Scenarios Modeling Scenarios - Discussion For modeling we simulated payment under various APR DRG payment scenarios for each claim in the analytical dataset. Per DHS direction, for evaluation purposes we used a standardized, simplified approach to payment system parameters in the models. All modeled scenarios included the following payment system parameters: APR DRG grouper version 35, the latest available grouper version APR DRG version 35 standard national weights and average lengths of stay Standardized DRG base rates consisting of statewide standardized amount, adjusted by hospital FFY 2018 Medicare IPPS wage index, solved for to achieve DHS' target expenditures under each scenario (budget neutral to current system expenditures in aggregate) Medicare-style cost-based outlier payment method with $30,000 fixed loss threshold and 80 percent marginal cost factor Medicare-style standard pro-rated transfer payment method 20 For evaluation purposes, we developed three different model scenarios using the above parameters. As previously mentioned, we recommend use of All-Patient Refined Diagnosis Related Groups, or APR DRGs (as specified in the Legislation mandating this study) and the use of 3M APR DRG standard national weights. While the modeled scenarios provide examples of other system parameters, finalization of these parameters will require additional modeling, stakeholder input and DHS decision making. As such, these modeled scenarios were designed as examples for evaluation of select system parameters and do not represent a final or recommended set of model parameters or rates. Should the Arkansas Legislature and DHS decide to move forward with implement a DRG payment methodology, then additional consideration and stakeholder discussions will be required to finalize the set of options that will comprise the full DRG pricing method for Arkansas Medicaid. The three modeled scenarios are as follows: Simulation 01 Model Using Current System Allowed Amount for DRG Funding Pool: DRG funding pool based on claim allowed amounts only ($581,815,376) as basis for the budget target. The baseline model does not include service line or population policy adjusters, and distributes payment across providers based on volume and case mix. The baseline model is not intended as a viable model for implementation, but rather as a starting point for evaluating where adjustments are needed in subsequent model scenarios. Simulation 02 Model Using Current System Allowed Amount Plus Supplemental s for DRG Funding Pool: DRG funding pool based on 20 Transfer claim identification uses patient status code/patient discharge status code values: 02, 05, 65, 66, 82, 85, 93, 94. Transfer claims are paid the lesser of: the full DRG payment or the pro-rated per diem payment calculated as: (hospital base DRG payment divided by the national length of stay) times (actual length of stay plus one day) Page 47 December 29, 2017

52 Arkansas DRG Conversion Plan combined claim allowed amounts ($581,815,376) and supplemental payments ($275,390,153) as basis for the budget target ($857,205,529). The baseline model does not include service line or population policy adjusters, and distributes payment across providers based on volume and case mix. The baseline model is not intended as a viable model for implementation, but rather as a starting point for evaluating where adjustments are needed in subsequent model scenarios. Simulation 03 Alternative Model Using Current System Allowed Amount Plus Supplemental s for DRG Funding Pool: Uses same combined claim allowed amount and supplemental payments as basis for DRG funding pool as Model 02 ($857,205,529). This alternative model introduces the following example policy adjuster factors applied to base DRG payments, determined to result in simulated DRG payments approximately equal to current system payments: 1.90 for normal newborn DRG claims 1.50 for neonatal DRG (NICU) claims 1.30 for all non-nicu, pediatric claims where the recipient was less than 18 years of age 1.0 factor for all other claims These modeled policy adjusters are for demonstration purposes; DHS may wish to explore additional, more-refined policy adjusters before finalizing system parameters. Table 9.2 shows the aggregate totals for each of the three model versions: Page 48 December 29, 2017

53 Arkansas DRG Conversion Plan Table 9.2 Model Totals Model Totals Model 01 with Claim s Only Model 02 with Claim and Supplemental s Model 03 Alternative with Claim and Supplemental s CY 2016 Admissions 96,114 96,114 96,114 State-Wide Base Rate $ 7, $ 11, $ 9, APR DRG v35 Case Mix Billed Charges Amount $ 2,355,440,778 $ 2,355,440,778 $ 2,355,440,778 Current System Claim Allowed Amounts $ 581,815,376 $ 581,815,376 $ 581,815,376 Gross Supplemental s $ 0 $ 275,390,153 $ 275,390,153 Total DRG Funding Pool $ 581,815,376 $ 857,205,529 $ 857,205,529 APR DRG $ 581,815,102 $ 857,205,122 $ 857,205,167 Estimated Change $ (274) $ (408) $ (363) APR DRG Outlier $ 87,935,363 $ 71,565,950 $ 68,475,867 APR DRG Outlier Percentage % 7.99% 9.3 Modeling Findings Modeling Findings - Discussion We discuss model findings and considerations as follows: Simulation Model 01 ( Model Using Current System Allowed Amount for DRG Funding Pool) demonstrates that without policy adjusters there are significant negative payment changes for key high Medicaid utilization providers and key service lines such as neonatal, normal newborn and pediatrics. This baseline model supports the use of policy adjusters to enhance payment beyond the standard APR DRG case mix to mitigate impacts and preserve access to care. For Simulation Model 02 ( Model Using Current System Allowed Amount Plus Supplemental s for DRG Funding Pool), the inclusion of supplemental payments yields a $4, increase in DRG base rate compared to Simulation Model 01. The higher DRG base rate results in a decrease in projected outlier payments as a Page 49 December 29, 2017

54 Arkansas DRG Conversion Plan percent of total payments, down from approximately 15 percent in Model 01 to 8 percent to model 02. Like Model 01, this baseline model supports the use of policy adjusters to enhance payment beyond the standard APR DRG case mix to mitigate impacts and preserve access to care. For Simulation 03 (Alternative Model Using Current System Allowed Amount Plus Supplemental s for DRG Funding Pool), the selected policy adjusters address the negative payment changes for the key service lines of neonatal, normal newborn and pediatrics found in Model 02. However, negative impacts for other key service lines and high Medicaid utilization providers remain. To make the targeted enhanced payments from policy adjusters budget neutral, the modeled standardized DRG base rate was reduced by $1, compared to Model 02. The current per diem-based system has a wide range in average payment per day across providers (approximately $350 - $4,200). Transition to standardized payment rates under DRGs would result in significant payment impacts for providers at both ends of the current per diem rate spectrum. However, a standardize payment system that minimizes the range of provider payment rates and reduces reliance on provider-specific costs would be more consistent with the guiding principles of equity and incentivizing efficiency. While policy adjusters in Model 03 demonstrate how DHS can mitigate the impact of key Medicaid service lines under APR DRGs, there are still significant negative impacts for certain high Medicaid utilization providers that will require additional consideration, modeling and adjustments. In addition, provider-level impacts from transitioning supplemental payments into the DRG payment system would necessitate significant changes for payment streams such as the assessment and IGT programs that optimize funding sources outside of state GRF. For government-owned providers, the state share of these supplemental payments is in large part funded by IGTs. For privatelyowned providers, the state share of these supplemental payments is in large part funded by hospital assessments. The IGTs and assessments would have to be re-determined in order to produce enough net positive impacts to make acceptable in the hospital community. We understand Arkansas Medicaid is planning to transition select populations to a Medicaid managed care program. Under Medicaid managed care, the flexibility for the Medicaid agencies to make supplemental payments to individual hospitals is significantly reduced. Within standard Medicaid rules (that is, without a Medicaid waiver), the state agency can only distribute supplemental payments directly to hospitals within Upper Limit (UPL) limits, which apply only to the fee-for-service program. As the size of the fee-for-service program decreases (with corresponding increase in Medicaid managed care), the amount of funds that can be distributed through supplemental payments also decreases. The only exceptions to this rule, short of negotiating a Medicaid demonstration waiver, are supplemental payments for Graduate Medical Education for the Disproportionate Share Hospital program. Shifting some or all of supplemental payment funding into claim-based payments through the DRG base rate or other claim payment addon is a potential method to distribute these funds to hospitals in a Medicaid managed care environment. CMS is currently emphasizing to Medicaid agencies that the majority of Medicaid Page 50 December 29, 2017

55 Arkansas DRG Conversion Plan reimbursements need to be tied to utilization of services. In particular, under a Medicaid managed care environment, CMS has stated that pass-through supplemental payments included in managed care capitation rates will be phased out over the next few years. Instead, moving supplemental payments in Arkansas currently funded through non-grf sources into per-claim payments is a permissible option to satisfy CMS s current policies and maintain funding levels. However, this method cannot guarantee the exact annual distribution of payments to each hospital. Given the federal limitations in place today and the planned movement toward implementation of the PASSE and other innovative service delivery models, Arkansas will need to develop strategies for repurposing traditional supplemental payment funding under alternative arrangements. The DRG conversion provides Arkansas with the opportunity to repurpose funds towards claim based payment. However, given Arkansas reliance upon non-grf funding sources (IGTs/ assessments), the State must find the balance between their policy objectives under transformational initiatives such as DRGs and managed care and the payment impacts to contributing providers (as the two may conflict). As Arkansas considers options, it is critical to have a strong understanding of provider payment impacts, federal limitations, and downstream impacts on funding sources. Page 51 December 29, 2017

56 Arkansas DRG Conversion Plan 10 DRG Implementation Fully implementing a CMS-approved DRG methodology under DHS FFS inpatient payment system will require several steps. This section describes key activities required for successful implementation and references estimated time requirements for each task. Implementation tasks are summarized in Figure 10.1 below. Generally, Tasks 1-3 can be conducted concurrently. Tasks 4-7, which can also be conducted concurrently, cannot begin until after Tasks 1 is complete. Figure 10.1 DRG Implementation Tasks 1 - Finalize APR DRG Method Design Approximately 3-6 months 2 - Evaluate Assessment and IGT Arrangements Approximately 6-9 months 3 - Stakeholder Engagement and Training Approximately 6-12 months 4 - Develop DRG Calculator 1 month 5 - Administrative Code/State Plan Changes and UPL Approximately 3 months 6 - Develop Access Monitoring Review Plan Approximately 3 months 7 - Determine MMIS Business Requirements and Make Required System Changes 6-12 months 10.1 Task 1 Finalize APR DRG Method Design Finalize APR DRG Model Initial DRG conversion review included in this report consisted of baseline modeling without (Simulation 01) and with (Simulation 02) supplemental payments as well as policy adjuster considerations (Simulation 03) using calendar year 2016 claims. Additional modeling should evaluate the impact of all policy considerations prior to finalizing a payment model, evaluating impact through incremental and iterative simulations. In addition, before finalizing the DRG model, DHS may wish to use more recent claims data Validate Comprehensiveness and Accuracy of Coding in Final Model Dataset Prior to finalizing a payment model, additional analyses surrounding documentation and coding may be warranted. Incomplete or inaccurate coding can have significant impacts on prospective rates and should be evaluated to ensure budget neutrality and equitable payments. As previously mentioned, initial evaluation of the model dataset identified several thousand claims where the delivery result (e.g. single live newborn) was reported as the primary diagnosis Page 52 December 29, 2017

57 Arkansas DRG Conversion Plan rather than main reason for the hospitalization (e.g. vaginal delivery). In this example, literal processing of these claims based on reported diagnosis code order would result an APR DRG unrelated to obstetrics services. Coding practices should be investigated in terms of both accuracy and completeness, comparing to national standards, comparable Medicaid agencies (external validity) and between providers and provider types (internal validity) Task 2 Evaluate Assessment and IGT Arrangements Transition to DRGs are likely to cause fiscal impacts that may require a new evaluation of the Access and UPL payments and the provider assessments and IGT arrangements that fund them. This will be particularly important should DHS decide to transition supplemental payment funding into the DRG system (as demonstrated in Simulations 02 and 03). Should DHS elect to include a portion of or all supplemental payment funding into the DRG system, a comprehensive review of the Access and UPL payment calculations is required including revisions to the assessment and IGT basis Task 3 Stakeholder Engagement and Training For the purposes of transparency and collection of input from the hospital community, a public stakeholder process is imperative. The type and frequency of stakeholder meetings vary by State and examples include: All-provider meeting(s): share key information such as timelines and key decisions Technical Advisory Group (TAG) meetings: solicitation of feedback during the iterative modeling process from a subset of stakeholders Hospital Association meetings: provide modeling update, solicitation of feedback and timeline status updates Legislative meeting(s): as needed, to highlight policies and budget implications After finalization of DRG payment method design, hospital and support staff training is required prior to implementation. The approach for training also varies among States, often including a combination of on-site ( live ) and virtual (e.g. WebEx) meeting options. Multiple training sessions may be warranted to ensure adequate opportunities for attendance among stakeholders Task 4 Develop DRG Calculator Upon completion of modeling, a DRG Calculator should be developed to demonstrate the price for a single inpatient claim. The DRG calculator is typically a Microsoft Excel based tool, and has proven extremely useful for all stakeholders, including the Medicaid Agency, hospitals, managed care plans and other impacted contractors (e.g. Medicaid Fiscal Agent, 3M). A DRG Calculator allows a user to enter specific claim fields (e.g. submitted charges, patient age, APR DRG) and after which, the tool will calculate claim payment, including documenting all incremental steps used to determine final payment. Note that the DRG Calculator does not group claims under APR DRGs (rather uses the APR DRG as an input for pricing); the user must separately group claims using 3M s APR DRG software to obtain the APR DRG assignment. Wisconsin 21 and Florida 22 are example Medicaid agencies that update their calculators annually and provide the tools through their public website Page 53 December 29, 2017

58 Arkansas DRG Conversion Plan 10.5 Task 5 - Administrative Code/State Plan Changes and UPL The inpatient reimbursement section of the Arkansas Administrative Code and State Plan will require significant revisions for an APR DRG reimbursement system to be approved by the State and CMS. A vendor can assist the Division with updating the Arkansas Administrative Code and filing a State Plan Amendment (SPA) for the new inpatient system, ensuring accuracy and completeness. CMS requires an inpatient Upper Limit (UPL) demonstration showing that Medicaid payments do not exceed payments under Medicare in order to approve a State Plan Amendment involving a major change in payment method. DHS or a vendor will need to develop an inpatient UPL demonstration for the first state fiscal year of the new APR DRG system, using simulated payments under the new system Task 6 Develop Access Monitoring Review Plan The major SPA change for DRGs will likely result in CMS requiring DHS to develop an Access Monitoring Review Plan (AMRP) to meet new CMS requirements to document whether Medicaid payments are sufficient to enlist providers to assure beneficiary access to covered care and services consistent with section 1902(a)(30)(A) of the Social Security Act Task 7 Determine MMIS Business Requirements and Make Required System Changes A transition from a per diem to a DRG reimbursement methodology requires significant changes to the pricing logic within the MMIS. A business requirements document for MMIS developers is critical to ensure that payments calculated within the MMIS accurately mimic the payment method design determined in Task 1. The Business Requirements Document (BRD) details the payment method design in terms that can be translated into more technical specifications, which identify where and how the MMIS software will be changed. In particular, the BRD documents the data elements, claim edits and formulas needed for calculation of DRG reimbursement. Thus, the document cannot be finalized until after the payment method design is completed. Also, the document should identify existing claim edits that are applicable to the per diem payment method, but are not needed under a DRG payment method. Once the business requirements document is complete, implementation of DRG pricing in an MMIS typically takes between six and twelve months. Changes needed in the MMIS include creation of an interface between the MMIS and DRG grouping software, which allows for a DRG code to be assigned to each claim during the process of adjudication. Generally, MMIS changes also include addition of new reference data, addition of logic used to determine the Medicaid allowed amount, and modification of reports and data extracts to include new data elements related to DRG pricing. Changes to an MMIS also usually involve changes to user manuals, billing manuals, claim suspense processing manuals, and call center response instructions. Although the MMIS changes will be significant, DHS s MMIS contractor recently went through APR DRG implementation in Wisconsin, demonstrating their ability to successfully implement this type of methodology. Page 54 December 29, 2017

59 Arkansas DRG Conversion Plan Appendix A: Summary Results of DRG Modelling for Arkansas Included in Appendix A are model summaries for each of the three model versions: Simulation 01 Model Using Current System Allowed Amount for DRG Funding Pool Simulation 02 Model Using Current System Allowed Amount Plus Supplemental s for DRG Funding Pool Simulation 03 Alternative Model Using Current System Allowed Amount Plus Supplemental s for DRG Funding Pool Page 55 December 29, 2017

60 Report A: Model Parameters Preliminary Medicaid APR DRG Analyses Arkansas Department of Human Services Arkansas DRG Conversion Plan Simulation 01 - Model Using Current System Allowed Amount for DRG Funding Pool Note: APR DRG Claims Dataset uses calendar year 2016 paid claims. Medicaid payments reflects full allowed amount (prior to TPL and/or copayments) and cost is estimated using Medicare cost-tocharge ratio times as submitted charges (uninflated). When included, gross supplemental payments allocated to each model claim based on charges (not net of IGT or tax). All claims grouped under APR DRG version 35 with 3M national weights. Simulation Parameters All Providers Comment Stays 96,114 State-Wide Base Rate $ 7, APR DRG v35 Case Mix Billed Amt $ 2,355,440,778 Est Cost (Billed CCR) $ 702,127,500 Allowed Amount $ 581,815,376 Equals sum of per diem payments for claims with date of discharge in CY 2016 Est. Gross Supplemental $ - Total Est Claim Pmt $ 581,815,376 Intention is budget neutrality APR DRG Simulation Pmt $ 581,815,102 Pmt Change $ (274) APR DRG Outlier Pmt $ 87,935,363 APR DRG Outlier Pct 15.1 Est Gross Supplemental Included in Budget Wage Index Adjustment of Base Rate DRG Policy Adjustor(s) Age Policy Adjustor(s) Provider Policy Adjustor(s) Documentation and Coding Adjustment No Yes No No No No Relative Weights APR v.35 National Transfer Policy Yes Discharge status codes: '02', '05', '65', '66', '82', '85', '93', '94' Outlier Policy Yes: $30,000 / 8 Medicare-like outlier policy: High side threshold (provider loss) and marginal cost (MC) percentage Charge Cap Policy No Page 56 December 29, 20017

61 Report B: Summary of Simulation by Service Line - Sorted by Percent Change Preliminary Medicaid APR DRG Analyses Arkansas Department of Human Services Arkansas DRG Conversion Plan Simulation 01 - Model Using Current System Allowed Amount for DRG Funding Pool Note: APR DRG Claims Dataset uses calendar year 2016 paid claims. Medicaid payments reflects full allowed amount (prior to TPL and/or copayments) and cost is estimated using Medicare cost-to-charge ratio times as submitted charges (uninflated). When included, gross supplemental payments allocated to each model claim based on charges (not net of IGT or tax). All claims grouped under APR DRG version 35 with 3M national weights. Service Line Stays Casemix Charges Estimated Cost Allowed Amount Gross Supplemental Proposed Change Percent Change Pay / Cost Pay / Cost A B C D E F G H = F + G I J = I - H K = J H L = H E M = I E N O = N I Circulatory Adult 3, $ 186,282,924 $ 46,152,874 $ 14,117,927 $ - $ 14,117,927 $ 40,593,415 $ 26,475, % 3 88% $ 1,836,648 5% Misc Adult 15, $ 582,905,305 $ 156,550,791 $ 68,675,645 $ - $ 68,675,645 $ 141,986,237 $ 73,310, % 44% 9 $ 8,331,623 6% Gastroent Adult 4, $ 132,562,673 $ 35,411,072 $ 18,202,717 $ - $ 18,202,717 $ 32,674,835 $ 14,472, $ 709,043 Resp Adult 3, $ 100,624,313 $ 27,237,552 $ 14,019,498 $ - $ 14,019,498 $ 23,608,121 $ 9,588,623 68% 5 87% $ 841,661 4% Rehab $ 125,239 $ 33,314 $ 22,850 $ - $ 22,850 $ 35,855 $ 13,005 57% 69% 108% $ - Burns $ 17,338,750 $ 7,695,418 $ 4,489,665 $ - $ 4,489,665 $ 6,105,184 $ 1,615,519 36% 58% 79% $ 3,610,935 59% Obstetrics 16, $ 252,009,455 $ 63,674,029 $ 38,979,245 $ - $ 38,979,245 $ 50,237,890 $ 11,258,645 29% 6 79% $ 161,956 Substance Abuse 1, $ 13,048,107 $ 3,422,118 $ 3,553,115 $ - $ 3,553,115 $ 3,417,639 $ (135,476) -4% 104% 10 $ - Misc Pediatric 6, $ 266,829,135 $ 101,270,801 $ 103,216,468 $ - $ 103,216,468 $ 80,686,780 $ (22,529,688) $ 27,057,435 34% Neonate 3, $ 317,668,157 $ 100,514,948 $ 105,568,888 $ - $ 105,568,888 $ 81,895,047 $ (23,673,841) % 8 $ 23,235,789 28% Mental Health Adult 6, $ 83,683,640 $ 22,662,377 $ 33,231,577 $ - $ 33,231,577 $ 25,413,799 $ (7,817,778) -24% 147% 11 $ 97,866 Transplant $ 22,405,053 $ 7,280,569 $ 8,643,736 $ - $ 8,643,736 $ 6,010,636 $ (2,633,099) % 83% $ 2,659,683 44% Resp Pediatric 3, $ 95,345,575 $ 37,923,007 $ 43,743,962 $ - $ 43,743,962 $ 28,942,628 $ (14,801,334) -34% 115% 76% $ 13,155,789 45% Normal Newborn 22, $ 113,047,651 $ 30,736,545 $ 46,011,027 $ - $ 46,011,027 $ 24,063,568 $ (21,947,459) -48% 15 78% $ 317,427 Mental Health Pediatric 9, $ 171,564,802 $ 61,562,086 $ 79,339,057 $ - $ 79,339,057 $ 36,143,468 $ (43,195,589) -54% 129% 59% $ 5,919,506 16% Total 96, $ 2,355,440,778 $ 702,127,500 $ 581,815,376 $ - $ 581,815,376 $ 581,815,102 $ (274) 83% 83% $ 87,935,363 15% Notes: C) Average APR DRG Weight. D) Billed Amount as submitted without inflation. E) Estimated cost using FFY 2017 Medicare IPPS PUF cost-to-charge ratio or most currently available HCRIS cost report data (July 2017 release) for providers not participating in IPPS. F) Full Medicaid Allowed Amount, not reflecting TPL or co-payments. Outlier Sim Outlier % of Pymt Page 57 December 29, 20017

62 Figure B Preliminary Medicaid APR DRG Analyses Arkansas Department of Human Services Arkansas DRG Conversion Plan Simulation 01 Pay-to-Cost Ratio by Service Line Pay-to-Cost Simulation Pay-to-Cost % Pay-to-Cost Ratio statewide average is 83%. 108% % 87% 69% % 79% 79% 6 104% % % 83% 115% 76% 78% 129% 59% % 2 Service Line Page 58 December 29, 20017

63 Figure B Preliminary Medicaid APR DRG Analyses Arkansas Department of Human Services Arkansas DRG Conversion Plan % Simulation 01 Percent Change in Reimbursement by Service Line 15 Percent Change in Reimbursement % 8 68% 57% 36% 29% -4% % -3-34% -48% -54% -10 Service Line Page 59 December 29, 20017

64 Report C: Summary of Simulation by APR DRG Severity of Illness - Sorted by Stays Preliminary Medicaid APR DRG Analyses Arkansas Department of Human Services Arkansas DRG Conversion Plan Simulation 01 - Model Using Current System Allowed Amount for DRG Funding Pool Note: APR DRG Claims Dataset uses calendar year 2016 paid claims. Medicaid payments reflects full allowed amount (prior to TPL and/or copayments) and cost is estimated using Medicare cost-to-charge ratio times as submitted charges (uninflated). When included, gross supplemental payments allocated to each model claim based on charges (not net of IGT or tax). All claims grouped under APR DRG version 35 with 3M national weights. SOI SOI Desc. Stays Casemix Charges Estimated Cost Allowed Amount Gross Supplemental Proposed Change Percent Change Pay / Cost Pay / Cost Outlier Sim Outlier % of Pymt A B C D E F G H = F + G I J = I - H K = J H L = H E M = I E N O = N I 1 Minor 47, $ 587,155,224 $ 164,759,295 $ 143,914,115 $ - $ 143,914,115 $ 131,480,607 $ (12,433,508) -9% 87% 8 $ 3,541,326 3% 2 Moderate 33, $ 689,549,669 $ 200,032,722 $ 177,078,774 $ - $ 177,078,774 $ 167,704,533 $ (9,374,241) -5% 89% 84% $ 8,138,663 5% 3 Major 11, $ 523,392,549 $ 156,534,361 $ 130,881,687 $ - $ 130,881,687 $ 131,311,967 $ 430,280 84% 84% $ 15,189, Extreme 3, $ 555,343,336 $ 180,801,122 $ 129,940,799 $ - $ 129,940,799 $ 151,317,994 $ 21,377,195 16% 7 84% $ 61,065,521 4 Total 96, $ 2,355,440,778 $ 702,127,500 $ 581,815,376 $ - $ 581,815,376 $ 581,815,102 $ (274) 83% 83% $ 87,935,363 15% Notes: A) Severity of illness (SOI) C) Average APR DRG Weight. D) Billed Amount as submitted without inflation. E) Estimated cost using FFY 2017 Medicare IPPS PUF cost-to-charge ratio or most currently available HCRIS cost report data (July 2017 release) for providers not participating in IPPS. F) Full Medicaid Allowed Amount, not reflecting TPL or co-payments. Page 60 December 29, 20017

65 Figure D Preliminary Medicaid APR DRG Analyses Arkansas Department of Human Services Arkansas DRG Conversion Plan Page 61 December 29, 20017